Assembly Standing Committee on Transportation

Sound check. Soundcheck. Soundcheck. Soundcheck. Soundcheck Sound check.

The Assembly Transportation Committee is called to order. Good afternoon, everyone. Welcome. The hearing room is open for watching in person, and this hearing can be watched from a live stream on the assembly's website. We seek to protect the rights of all who participate in the legislative process so that we can have effective deliberation and decisions on the critical issues facing California.

In order to facilitate the goal of hearing as much from the public within the limits of our time, we will not permit conduct that disrupts, disturbs, or otherwise impedes the orderly conduct of legislative proceedings. We will not accept disruptive behavior or behavior that incites or threatens violence. We encourage the public to provide written testimony by sending an email to [email protected]. Once again, that is [email protected]. This email address is also posted on our website.

Staff will be reading all comments, and we can have follow-up conversations as needed. Please note that any written testimony submitted to the committee is considered public comment and may be read into the record or reprinted. At the end of this hearing, we will allow for one minute each, and I will remind you that it may be less depending on the amount of public comments we receive. One minute each for each witness, from the public to speak regarding the project or regarding this hearing.

The purpose of today's hearing is discuss the regulations finalized on 04/28/2026 for the testing and deployment of autonomous vehicles on California's roads.

During our hearing today, the Department of Motor Vehicles and the California Highway Patrol will make brief opening remarks and provide a overview of the finalized regulations. We will also hear from autonomous vehicle industry partners, laborers, consumer groups, first responders, and researchers on the new autonomous vehicle regulations. These regulations are a much needed update to the regulations adopted in 2018, which have been inadequate in recent years to addressing the growing landscape of autonomous vehicles. Prior DMV regulations stopped collecting crash data after autonomous vehicles received deployment permits. No test existed for graduating from a driver's driver testing permit to a driverless testing deployment.

No data was being collected on vehicle immobilizations and interactions with first responders have been inadequate. The December blackout in San Francisco, which resulted in over a thousand AVs stalled for at least two minutes and hour long wait times for first responders to contact the AV company have shown that both government and the AV industry have a long way to go, and the new regulations are far more comprehensive than the previous regulations.

At the same time, the legislature needs confidence in the DMV to regulate AVs as the department has now authorized 80,000 pound trucks to test and deploy on California's roads without a human operator. I'm look forward to I look forward to learning today, from our panelists about the new regulations and if the stakeholders think that they strike the right balance between protecting public safety while also providing a path forward for the driverless innovation being developed right here in California.

Members will be less limited to two minutes each of opening remarks since we have a full agenda, and I definitely wanna allow for robust discussion while we have panelists in front of us. With that, I look to, the members that are present to see if they would like, to to share any opening remarks. Assemblymember doctor Corey Jaskensen.

Thank you very much, madam chair, and thank you for your wisdom and making sure that there's plenty of dialogue when it comes to this issue. As you know, technology is moving fast. And as our communities continue to adapt to it, we have to make sure that we're leading with the best information possible, including what trade offs are we making when it comes to autonomous vehicles both good or bad. Does autonomous vehicles help to make our streets safer or do they not?

Do autonomous vehicles, are they a disruptor when it comes to other to our workforce? We have to make sure that we're taking a deep dive into these issues. I've certainly been waiting to hear from the DMV when it comes to their assessments, and I look forward to hearing what their assessments are when it comes to this technology. But I'm also looking forward to seeing, you know, what does this present to our future sectors in terms of transportation, in terms of goods movement, and in terms of all the other things that we're relying on, as well. So thank you very much for your leadership on this issue and looking forward to this robust conversation.

Thank you, doctor Jackson. With that, we will move on to our opening remarks. We'll begin our hearing, and I'd like to invite now Bernard Soriano and Miguel Acosta with DMV and assistant chief Ty Meeks and Lieutenant Commander Dave Falson with the highway, California Highway Patrol. They're gonna give our opening remarks and speak about the regulations.

Thank you, doctor Jackson. With that, we will move on to our opening remarks. We'll begin our hearing, and I'd like to invite now Bernard Soriano and Miguel Acosta with DMV and assistant chief Ty Meeks and Lieutenant Commander Dave Falson with the highway, California Highway Patrol. They're gonna give our opening remarks and speak about the regulations.

You can begin at your convenience.

Make sure to press the button. As a reminder for this opening remarks, you have ten minutes combined.

Thank you, madam chair. Good afternoon, madam chair and committee members. Thank you for the opportunity to testify. California continues to be a leader nationally and internationally in the regulation of autonomous vehicles. The California DMV's balanced approach, encouraging innovation while ensuring public safety is guided by a comprehensive regulatory framework established rules issued in 2014, 2018, 2019, and now in 2026.

This framework is among the most rigorous, structured, and transparent in The United States, emphasizing safety, accountability, and strong coordination with law enforcement and emergency responders. California's approach remains aligned with the Federal Government's model in which manufacturers self certify compliance with federal motor vehicle safety standards, while states oversee driver licensing, vehicle registration, insurance, and operational safety. The department currently administers three types of autonomous vehicle permits, testing with the safety driver, driverless testing, and deployment.

Today, 27 manufacturers hold driverless driver testing permits, six hold driverless testing permits, and three are authorized for deployment. The new regulations became effective on 04/28/2026.

And throughout the rule making process, the department engaged extensively with local partners, including transportation agencies and first responder departments in Los Angeles, Santa Monica, and San Francisco. And we convened two statewide first responder roundtables to ensure operational concerns were fully incorporated. California remains the only state with a formal phase permitting structure that governs the full life cycle of AV operations from driver testing to driverless testing, driverless operation, and commercial deployment.

This tiered system requires manufacturers to demonstrate readiness at each stage, including submitting safety cases, validating operational design domains, and providing comprehensive data reporting. For example, manufacturers must now test 50,000 miles for light duty AVs and half a million miles for heavy duty AVs, and they must submit a structured safety case demonstrating the safety of the vehicle's hardware, software, and operations.

These new regulations establish a path for deployment for testing and deployment of heavy duty autonomous vehicles and strengthen strengthens the DMV safety oversight. They removed the previous prohibition on autonomous vehicles with gross vehicle weight rating of a 10,000 pounds or more, and heavy duty autonomous vehicles must still comply with all applicable state and federal commercial motor vehicle requirements, including the CHP's weigh stations and inspections.

Manufacturers of heavy duty AVs like operators of human driven commercial motor vehicles must use roadways that are legal for the size, weight, and loading of the vehicle. AV manufacturers are prohibited from carrying oversized loads. They are also prohibited from transporting hazardous materials.

Additionally, the regulations allow for AVs up to 14,000 pounds gross vehicle weight to be operated by public by a public entity or university for passenger transit. These passenger operations must provide the department with any terminal evaluations from the CHP, and the department may restrict or suspend operations if the Federal Transit Administration issues a directive restriction or prohibition. At this point, I'll turn it over to Miguel Acosta, also from the DMV, who will speak to the reporting requirements of these new regulations. Miguel?

The regulations enhance reporting requirements by mandating more detailed and frequent submissions, including reports on vehicle immobilizations, hard braking events, system failures, and federally aligned crash reporting. Previous regulations required crash and annual disengagement reporting. The updated requirements provide more frequent data, enabling the department to better review incidents and assess manufacturer applications. In addition to the department suspension authority, the regulations formalize the department's request for information process, allowing it to seek specific details regarding the root cause of incidents and any identified remediation.

Enforcement tools are also expanded. The department may impose specific restrictions on a manufacturer's operations based on an incident, such as limits on fleet size, geographic area, time of day, or roadway types. The regulations also implement Assembly Bill 1777 by requiring AVs to respond appropriately to emergency geofence messages, comply with law enforcement direction, and support first responders when an override system is present.

The rulemaking further creates a formal process for law enforcement to notify both the department and manufacturers when an AV is involved in a moving violation. The regulations establish standards for remote operations personnel, including qualifications, training, and permitting requirements for remote drivers and assistance.

Manufacturers must provide details on how request to remote assistance are assigned and how staffing levels are determined. Each remote agent must only be assigned requests for which they are trained and certified. AV manufacturers must also demonstrate robust remote operations communication systems and provide the department with information on response times and latency thresholds. Remote drivers who can perform the driving task must hold a valid driver license for the vehicle type.

They must maintain clean driving records and are subject to the same regulatory requirements as autonomous vehicle test drivers.

The department maintains strong oversight through ongoing incident review, data reporting, and post permitting enforcement. Since 2014, the department has issued three permit revocations, 14 suspensions, including the 2023 suspension of cruise driverless testing and deployment permits in San Francisco and the 2021 suspension of Pony.ai following a system failure related crash in Fremont, California. California's regulatory framework supports the continued development of autonomous technology while prioritizing roadway safety, compliance with state law, and safe interactions with other road users.

Thank you for this opportunity today, and we welcome any questions. I'll turn it over to my colleagues at the highway patrol.

Good afternoon, madam chair. Tye Meeks with the California Highway Patrol Office of Special Representatives. With me is Lieutenant David Fossen, the commander of our collision investigation unit. We're grateful for the opportunity to to be here, with our counterparts, of the DMV. I can assure you that both DMV and the CHP have been laser focused throughout this process to make sure that both public safety and transportation safety has been at the forefront of these regulations.

And that through collaborate the collaborative process, law enforcement and AV industry, have worked together to come up with what these regulations look like today. A multitude of public hearings have been held in an effort to address concerns and show transparency in the process. We are confident that the regulations provide a mechanism for law enforcement agencies, such as the highway patrol and local, departments to hold AVs accountable when violations of the vehicle code are observed and when traffic violations, of concern are realized.

The CHP's commercial vehicle section provided, subject matter expertise to the DMV regarding, commercial vehicle operations and enforcement as it pertains to, commercial vehicles and both federal and state laws. The department is ready for commercial vehicles to come through enforcement facilities in line with the current, DMV regulations, as outlined with test drivers.

The CHP crash investigation unit provided assistance to the DMV reviewing and approving all of industry's first responder interaction interaction plans that are accessible to all first responders and each AV, through each AV manufacturer's website. Some companies have gone so far as to place a QR code on the side window next to the side mirror to make it, easily accessible to, any first responder, any, individual showing showing up at the scene.

Additionally, the CHP has attended nationwide AV workshops and participated in, AV working groups, which were attended by local law enforcement partners here in California and throughout the nation. These practices are utilized and improved have helped us to improve the current regulations providing California with some of the most comprehensive regulations in the nation. These regulations are an example of our commitment to both hear and address concerns that have been raised by all.

The regulations make sure that California remains true to its calling, card of being a leader and the standard bearer for in the nation when it comes to innovation. We provided presentations in person and online to several local agencies and city officials on how to utilize the notice of noncompliance for autonomous vehicles. Santa Monica, these are just a few. Santa Monica, San Francisco, Mountain View, San Jose, Beverly Hills, Los Angeles, and Riverside.

An online training video has been, developed and is posted on YouTube on and the DMV's website for quick access on how to utilize, complete, and, what to do with the notice of noncompliance for, autonomous vehicles when it's filled out.

Both departments have prepared information bulletins and memos outlining the process on how to requisition, the notice of, autonomous vehicle noncompliance, and step by step instructions on how to fill it out and what to do when it is once completed. Chair and members, thank you for, the opportunity to testify today. As previously noted, California leads the nation when it comes to to innovation. But we have been equally clear that innovation cannot come at the expense of public or transportation safety. The autonomous vehicle regulations reflect that balance.

They are grounded in, data, strengthened by continuous oversight and built on the principle that every operator, whether human or autonomous or automated, must be held fully accountable on our roads. Our commitment is unwavering to ensure California remains a safe place to live, work, and travel for those who are visiting this great state and those who live within the state. We're happy to answer any questions as far as it comes to the enforcement part, portion of autonomous vehicles.

Thank you. Appreciate that. With that, I'll look to before I move into any questions, I'll look to members of this committee if they had any questions. Doctor Jackson?

From the data that you've all seen, I would love to hear from all of you and on this is I mean, obviously, is the question about safety and which type of driver is the safest driver. I must admit, I'm not one of them. But do you yeah. Good to know. If you see me, you're in danger.

I have to agree with you, doctor Jackson, for myself. Not for you, for myself as well. We're we're assertive drivers probably is a nice word.

I was born in Los Angeles. That's all I can say. Are we finding that autonomous vehicles are yielding safer outcomes for especially in our urban areas at this time? Do we have enough data to make that claim yet?

Yeah. I can start and Miguel can can fill in. You know, certainly, we have a plethora of data of human drivers. Right? But if you if you try and make a comparison between human drivers and, autonomous vehicles, the data itself, there's there's a big disparity.

So I think it's It's still too early to definitively say. But from what we have seen, the total number of collisions that were reported, at least between 2014 and 2026, were less than a thousand. These were with AVs and these would be testing with a driver, also testing without a driver. The majority of them are with with a driver on board.

So with that small dataset, you can also see that, the number of collisions that were reported based on and compared to the number of miles that were driven were much smaller than the total number of miles that were driven and the number of collisions that happened with with humans.

So it's still early to to definitively say, but it seems to be trending in that right direction. Here we go.

And just to expand on that some further, I mean, I as as what Bernard said is is right that the the data that we had was really focused on testing. And the regulations that we have now are gonna expand that crash reporting to just testing and deployment. But back to the testing data that we do have, of those about 983 collisions that we saw during testing, Forty percent or so were were were the AV was rear ended by a human driver. So it's where it's where the human driver and maybe for a variety of reasons. Right?

It could be the the driver was not paying attention. But the AV was at a stoplight, and and the AV was rear ended by a by a human driver. And then other than other data to notice, about sixty eight percent of crashes, this is a during testing, were resulted in minor damage. So these were very minor types of collisions. So there were no no fatalities reported during that time.

But again, this is a small dataset, but this is the data that we have available. But I think it points to that in this new regulations, we're gonna have much more data. We're gonna continue to learn from what that data provides.

From the highway patrol's perspective, we can attest that here in California, we have not recognized or seen an incident where an autonomous vehicle in levels three or four autonomous mode have been involved in a fatal traffic collision. So from our data and our our recollection.

And, I mean, obviously, there's some of the most glaring complaints is the idea of when the technology goes wrong and it begins to stall, and then it's just holding everyone up including our first responders. Are are we getting better at solving that issue when it comes to first responders and they're trying to get to where they need to go and how autonomous vehicles have been reacting to that?

I've noticed in your that you're getting a little bit better on your requirements in terms of how to make sure the technology is talking to first responders. Can you elaborate more on that?

Well, maybe I'll start and then I'll see if CHP wants to do it right. But that that's absolutely right that we need more of those requirements for the manufacturers to report those those immobilizations and those interactions with first responders.

As mentioned in our opening statement, prior to the regulations, we have fostered relationships with local governments such as San Francisco and others to to really try to understand these incidents that are occurring on public roads, and they really informed our regulations and the need for this data. And so the new regulations will give us that information regarding immobilization's interactions with first responders.

But in regards to the incidents that we do have presented to us, we work directly with those manufacturers to understand why that is occurring and how they're particularly gonna be remediating those types of issues.

So it's not only receiving the data, but it's also working with the manufacturer to see how those issues are ameliorated.

There's an expectation that the autonomous vehicles will obey the rules of the road and that when confronted by emergency vehicles that they will yield appropriately. We have seen issues in the past, and we I believe we will continue to see issues. Just as we do with human drivers, we see it with autonomous vehicles. However, the issues are are different.

Every every situation has one element that may be different from the previous one, and that's where, as Miguel mentioned, with CHP and DMV working together, working through with the manufacturers that are out there, we're able to to address those issues, hold them accountable for it so that some of the issues we've seen maybe a year or two ago, we don't see today.

However, we're seeing other issues that come forward, and that's where we have that open working relationship with them with the, with the inclusion now of the notice of AV noncompliance. This is a way for us now statewide to have the eyes of law enforcement officers to to directly report to the DMV of the issues that we have. Prior to that, it was usually issues from from other people, lay citizens out there, other people who see it, where it may or may not be be an issue.

Also, through our travels, we've seen that even in cities where CHP does not have primary jurisdiction, we do see that there are challenging roadway designs that do make it hard for for these AVs as well to to figure out where to pick up, where to drop off these people. And then as events happen, and then through also the the introduction of the of our ability to put out the the restrictions of locations, the geofence locations that'll we believe will help reduce these issues as well.

Okay. Thank you very much, madam chair.

Thank you. Okay. You got the question looking. Vice chair Davies.

Thank you, madam speaker or madam chair, I should say. Just a couple quick questions. With with the vehicles, what is the max speed they can actually go?

Well, I mean, the the vehicles themselves need to adhere to to the rules of the road, so whatever the posted speed limit is for the particular operating area. But, really, that that varies from the operational area that the that the AVs are are in. Like, if they're providing, let's say, delivery services in a neighborhood, they would have to adhere to the 25 mile per hour low lower. But if they're on the highways, right, freeway off speeds, they could travel up to that that speed.

Question. I mean, obviously, sometimes, there could be something ahead where you have to, as a car, speed up or you could get yourself into an accident if you see something, you know, a car on the side. Does it have that ability to do that then if it would need to and it couldn't slow down to avoid an accident?

So the regulations do provide for that that the the AV has to follow the rules of the road or the the vehicle code in except for instances, like you just mentioned, in instances where it has to respond to situations for for safety, for the safe operation. And obviously, there are edge cases like that. But I don't know, Dave, if you wanted to respond. Or

And then, I guess, my other concern, again, is we're talking about safety. Our Caltrans workers out there working on the freeway, and we already know that it's a very dangerous job when they're out there. And we are now finally getting some cameras to have, you know, hopefully, slow people down. Is that a concern as well? You know, they stay do they stay in the right right lane in the slow lane?

I'm just curious how they move back and forth.

Well, the AVs are if they're operating in the freeways or the that is their operational design domain, then, yes, they absolutely have to adhere to Sure. Rules of the road and the vehicle go, and that includes construction zones or yielding to to our Caltrans workers who are who are on the roadways and and or and law enforcement as well.

Thank you.

Thank you. And to one of your questions, we will have some of our operators on the panel. And so I would encourage to reask too when you have other other folks on the panel as well. So I have a few questions. One of the things we have some of this has relates to legislation that is in currently under review.

So under the DMV regulations, a remote driver and a remote assistant are required to have a valid driver's license. So is a driver's license that's issued by a foreign government considered a valid driver's license?

Yes. Under the regulations, that would be correct. Now, but there's a distinction between a remote assistant and a remote driver. And we made this in particular based upon the function that they provide. So remote assistant does not do any of the driving task, so they're not performing any of the driving responsibilities.

And so the requirements for remote assistant, they have to have a valid license. And that is so that they can have basic operational understanding of the rules of the road. Now remote driver is a little bit different. A remote driver is one that actually can perform the driving task from a remote location. And we have much more robust requirements for that particular remote operator, and they must have a valid driver license for the type of vehicle that they're operating.

And if it's a commercial motor vehicle here in United States, they would have to have a commercial driver license here that's issued here.

That's issued here. So if you had a foreign driver's license and you're a remote driver or operator, then you would not be eligible to drive a remote drive, a commercial vehicle without a US issued commercial driver's license. So anywhere in The US though. Right?

Correct. You would need a an approved CDL or commercial driver license to operate that vehicle as a remote driver. Yes. So it's much more stringent license requirements for the remote drivers. Because, again, they're the ones that are actually performing that driving test.

They're moving that vehicle remotely from a separate location.

Okay. Were you gonna add to that at all or no? Okay. Alright. And so wanted to ask about the regulatory actions as it relates to getting a notice of noncompliance.

And so do you all have, DMV, an objective standard for when to take action, based on the receipt of a notice of noncompliance?

Well, I think that one of the things that we're taking a look at is, every circumstance is going to be a little bit different. There's gonna be a variety of facts associated with that particular incident.

And once we receive a notice of noncompliance from a peace officer here in in California, we're going to investigate every single incident and determine whether or not, number one, is there any kind of imminent danger, imminent hazard that we need to review more carefully and take more immediate type of an action, one that is going to provide severe injury or or damage to property.

But, really, the standard we're gonna be using is we're gonna we're gonna be taking a look at all the facts associated with that particular incident and whether or not the manufacturer how they're gonna actually address or remediate that particular incident, and that's how we're gonna be evaluating.

Okay. Thank you. And then in that same vein of the notice of noncompliance, you all determined that it was better to place the reporting requirement on DMV versus not the entity that issued the noncompliance. So what was the reasoning piece behind that?

So the the reporting, so there's it's twofold. So the

I'm sorry. The operator. I'm sorry. Yeah. I messed that up.

Thank you.

So the main

But clean clean clean up your clean up my question and your answer. Okay.

So, yeah, the the manufacturer is responsible for providing us with that, citation so that we can start investigating that right away. Now if a law enforcement officer indicates on the form that it's a priority review, So these are these are incidents that where the officer in their professional judgment determines that this is something that we need to take immediate action review. The manufacturer has twenty four hours to send that to us.

So we also appreciate that there may be situations where there may be, there may be some kind of delay in that. So we're also requiring that the the law enforcement officer, the agency, also submits that information to us.

So I think the key is is that we wanna be able to act quickly on those incidents. And so we've actually required in regulation that both the law enforcement agency and the manufacturer provides it to us.

Okay. Thank you. This is to CHP, and this is really based on the fact that we just had the incident in December that, you know, made, national actually, probably global news. And so if an AV experiences a critical detachment or software freeze that blocks a lane on an interstate, what specific authority and training do CHP officers have to manually move or override the vehicle? And including in that, are you seeing, manufacturers responding within an acceptable time frame to clear these blockages?

So every autonomously driven vehicle is required to have a first responder interaction plan. In that plan, it does dictate how a first responder can enter the vehicle and take over manual control of that vehicle. That is not our first choice. However, as law enforcement, we sometimes have to get into a vehicle to move it to help with the expeditious flow of traffic. But keep in mind, where CHF's jurisdictions at many times on the freeways.

We wanna remove vehicles as quickly as we can out there. On surface streets, sometimes we have more time on our side. Sometimes moving the leaving the vehicle in place may be the better option. So I can't speak to the the time delays on surface streets is where we've seen more of this issue happen, But I can speak to a few incidents that we've had with CHP where officers were able to take over man out of control.

And other times, they were able to speak through the speakers on the car to the remote assistance, and they were able to move the vehicle to a safer location.

There is a bit of a time delay for that. Obviously, first responders especially respond if we respond to emergency incident, we may not have time to get out of a car to move that vehicle. And that's the that's the biggest frustration that we felt especially amongst the fire departments out there is we don't have the time resources to do that and we have to work our way around those vehicles many times.

However, just the the introduction of being able to move the vehicles, we have seen a a marked improvement in that. But that's just speaking from the CHP's perspective, not from the local PDs.

Thank you. Oh, did you wanna add?

Yeah. If I could add, I know the December incident was brought up earlier on and, what I can tell you is we received many of those calls dispatch wise. We transferred those to San Francisco PD, but we did have a unit that came across one of those, the vehicles that were stopped in the lanes. The officers approached the vehicle. They called the, the operator.

They called the the emergency line, and the car was moved within a minute. So from the time that they showed up to the time that, that car was cleared off to the right off to the curb, it was two minutes, for that call specifically. Now, that was just our experience.

Yeah. Thank you. I appreciate that. Last question for CHP and I'll give it to my the rest of my committee to make sure if there are any other questions for this panel, let me know. And so, you all operate CHP operates inspection facilities and weigh stations across our state.

And so, now that we're transitioning to uncrewed commercial, AVs, is that that transition, will that alter the standard inspection protocols? And then what, procedures would do you have in place if a AV truck bypasses or fails to properly interface with the weigh station? What is that, you know, immediate enforcement mechanism?

So currently, right now, even with human operators, we have, what's called, a bypass program, but we also have officers who are sitting at these facilities who are watching traffic as they come, the commercial traffic. If those commercial vehicles bypass the facility, whether it's human operator or, automated, they take a take action enforcement action. They'll make an enforcement stop, and they will take the appropriate enforcement action once they make that stop.

So that would be either the notice of autonomous vehicle noncompliance, or it would be, if it's a human operator, a citation of notice to appear. We are prepared and ready for autonomous vehicles to come through those facilities with those test drivers.

We are not currently prepared for full autonomy mode to go through those facilities. There are some things that still need to be worked out as far as locations to have them pull off and and the proper inspections to be conducted there.

Okay. But they will be held to the same standard as those being driven by a human operator.

Correct. The standard does not change.

Thank you. Alright. Double checking with members of this committee. I mean, sorry. Yeah.

Thank you so much, for your time. I appreciate it. We now be moving on to our, first panel. This was our opening remarks from DMV and, CHP regarding the new regulations. Now I would like to invite Elise, Sanguiette with the consumer attorneys and Renee Gibson with the Autonomous Vehicle Industry Association.

This will be our panel related to data collection and enforcement tools granted by the new AV regulations. You may begin at your convenience.

Thank you, madam chair. Oh, and real quick.

As a reminder, you each have your own five minutes. Yes.

Thank you so much. Good afternoon, madam chair and members. My name is Elise Sanquinetti. I am a past president of the Consumer Attorneys of California and a past president of the American Association for Justice, which is the national group. My experience includes working on legislation, regulation, and litigation involving autonomous vehicles for over the past for about the past decade, both in California and nationally.

I wanna thank you for the opportunity to be here to discuss the Department of Motor Vehicles' recently adopted autonomous vehicle, regulations and specifically their relationship to AB 3061, which was authored by assembly member Haney in 2024, which we cosponsored with the Teamsters and the Consumer Federation of California. When AB 3061 was introduced, one of its primary objectives was to ensure that California's autonomous vehicle oversight framework kept pace with rapidly evolving technology.

At the time, the DMV's autonomous vehicle regulations had remained largely unchanged, as the chair had mentioned, since they were adopted in 2018 despite significant changes in the autonomous vehicle deployment, expanded testing, and increased public interaction, and driverless vehicles on California roadways. The largest flaw that was, that was the d m was that the DMV neither collected nor reported data once an autonomous vehicle permit holder shifts from testing to a full deployment permit.

Cities have experienced, driverless vehicles shutting down, blocking intersections, causing gridlock, and even at times obstructing emergency vehicles as we've previously discussed.

As California serves as a test case for AVs, there is a public interest in the highest level of transparency. CAOC's position has remained that the that the need for innovation must be met with public transparency and a priority of safety. In addition to these incidents not only being yep. Not only being reported to DMV on oh, I'm sorry. On 10/02/2023, an autonomous cruise vehicle struck a pedestrian in San Francisco that was first struck by a human operated vehicle.

Cruise was not required to report this collision to DMV because it was operating on a deployment permit. AB 3061 sought to modernize California's approach by proposing reporting requirements for collisions, traffic, violations, disengagements, and other incidents involving autonomous vehicles operating on the public roadways. The goal was to ensure that regulators, policymakers, local governments, and the public have access to meaningful information about how these vehicles perform in real world conditions. Transparency and accountability are critical regarding this rapidly evolving technology.

I can speak as to civil litigation, obtaining critical data and other evidence of the causes of incidents prevents presents a significant challenge. Even in instances when someone has been seriously injured or killed, it takes years of discovery battles with companies. And I can speak to Tesla specifically, although they're not deployed here in California. It takes years to gather information within the company's control. The same is true for litigation against autonomous vehicle companies. This lack of transparency and accountability are an unnecessary drain on severely injured people, surviving families, and the courts.

We were pleased to see that the DMV recently adopted regulations that address many of the same concerns raised by AB 3061. Both the bill and the regulations recognize the importance of collision reporting, traffic law compliance, disengagement reporting, and increased oversight of autonomous vehicle operations. In that respect, there is significant overlap between the policy objectives of AB 3061 and the final regulations.

I'm sure I do want to address specifically trucks, although I know that, that there will be some discussion later panels on this issue. But I just want to remind the members that, that in instances where there are collisions, when we're talking about heavy duty vehicles, they do significantly more damage than collisions involving light duty autos.

And that is something to just keep in mind when we're talking about safety of our roadways. The discussion surrounding AB 3061 highlighted concerns, again, regarding transparency and accountability. Many of those same concerns are reflected in these regulations that were ultimately adopted. However and this is a big however, I want to urge that the public have access to the information that is collected, and that is not And

I'm so sorry.

I have to have you

Go ahead.

That is not currently, identified in these regulations. The public have access to that information, and that is critically important. So with that, I thank you for, for noting my time. Thank you from having me here, and I'll remain for any questions that the members have.

Thank you. Appreciate that.

Go ahead.

Thank you, madam chair, members of the committee. My name is Renee Gibson, and I'm the vice president of government affairs for the Autonomous Vehicle Industry Association, which represents the world's leading AV companies operating in California and across the country. AV companies have collectively driven more than 360,000,000 autonomous miles across The United States. Those miles represent years of testing, validation, and continuous improvement. California has long been at the forefront of AV development and regulation, but AVs do not appear on California roads overnight.

They are the product of a long and highly regulated process and has taken nearly fifteen years to reach this point. Over a decade ago, the DMV adopted the nation's most stringent regulations for AVs. After multiple workshops, hearings, and comment periods between 2023 and 2026, the DMV adopted revised rules, which set an even higher bar for AV companies.

The regulations established strict permit requirements for AV companies to even qualify for testing, ongoing data reporting requirements during operation, and broad authority for the for the DMV to restrict or suspend operations. No other state approaches AV policy with this level of permitting, reporting, and enforcement.

The DMV exercises oversight at every stage of AV operations. There is a progression through multiple permit stages designed to demonstrate operational readiness. It starts with driver testing, then driverless testing, and then companies are eligible for deployment. Before advancing to each stage, companies must meet specific mileage thresholds. Light duty AVs generally must test 50,000 autonomous miles, while heavy duty AVs must test at least 500,000 autonomous miles.

At each step, companies must submit detailed safety case information about vehicle design, the AV's operating conditions, testing and validation, remote operations, and first responder interactions, just to name a few. The regulations also require notification to local agencies, submission of detailed first responder interaction plans, and coordination with emergency personnel within an AV's operating area. This just touches the surface of the requirements, and I'll focus on a few specific aspects of the regulations.

With respect to data reporting and collection, the new regulations expand California's already robust transparency and reporting requirements. First, collisions must be reported consistently with long standing federal collision reporting requirements for AVs.

This means that companies have to submit the same information to the DMV that is submitted to NHTSA, the National Highway Traffic Safety Administration, on the same timeline that NHTSA requires. To be clear, the reporting threshold for that is very low, and companies have to report collisions regardless of fault. But the DMV's requirements go well beyond collision reporting now. Companies must also report system failures, vehicle immobilizations, harsh breaking events, notices of noncompliance, and vehicle miles traveled.

These reports must be provided monthly during testing and quarterly during deployment. Because these reporting not types are novel, I'll go in a little more detail on a couple of them. For example, immobilizations must be reported whenever a driverless AV is stopped in an active travel lane, cannot continue with the driving task, and must either be retrieved or removed by a human driver or a remote driver. Braking events must be reported during testing whether whenever an AV produces a specific speed decrease that is defined specifically in the regulations.

Finally, starting next month, companies must report any notices of noncompliance issued by traffic officials for a violation of state or local local traffic laws. These notices must be provided to the DMV within seventy two hours or within twenty four hours if the official indicates a need for priority review. This concept is designed to provide enhanced visibility into traffic enforcement scenarios. These reporting requirements do not exist elsewhere in the country and will provide DMV with unprecedented visibility into AV operations.

This visibility is coupled with substantial enforcement authority. The DMV has long been able to deny permit applications and suspend or revoke permits. Under the new rules, the DMV can also impose restrictions like reducing fleet size, limiting where and when vehicles can operate, or requiring a test driver to be present. And if the DMV determines that there is an imminent hazard, it may immediately suspend, revoke, or restrict a permit.

The combination of extensive reporting obligations and broad enforcement authority gives California regulators clear real time visibility into AV operations and the tools to intervene if it's if safety concerns arise.

The AV industry agrees with the intent of these regulations to provide safe and transparent operations on California roads. We take these requirements seriously and are spending significant time and effort on compliance. We look forward to continuing to support safer roads, mobility access, and supplying solutions to the state of California. Again madam chair, thank you for the opportunity to speak today.

Thank you, thank you to you both. Now I'll turn it over to members of this committee to ask any questions starting with assembly member Hart.

Well, thank you both for your testimony. Miss Sanguinetti, you mentioned that oftentimes, it can take years to get information from the AV manufacturers for court cases. And then miss Gibson mentioned that there are new DMV regulations that require additional reporting. Is that directly gonna work to solve the problem that you had raised in your testimony?

That's an excellent question. It it would if the information that's being collected by the DMV is publicly available. If if it is not publicly available, then the same battles will exist in the courts to obtain the information that's necessary.

So that's the question for the panel that that left us, but they'll be back, right, I think

Thank you.

At some point. Thank you for that. Miss Gibson, do you believe that this is gonna solve this court issue, or do you think there's still a gap?

I don't think there's a gap. Thanks for the question. I think, you know, there is a reason that the system exists today. There still needs to be investigations done. It needs to be fact that fact patterns that need to be explored.

You know, I think that I think that what is definitely true is that California has the most robust reporting requirements in the country, and that is gonna tell us a lot of information once it's all it's all in in effect, which is very soon.

Do you acknowledge that the, there was a gap previously and that this additional regulations are necessary to provide plaintiffs with the information they need in a court case?

I think the industry acknowledges that transparency is important. I think we recognize that we have to we are asking a lot of the of the public, right, to trust the technology. And and and showing our safety data is gonna be helpful in in earning that trust.

Okay. Thank you.

And just to note, the limited data that was collected in the past was made available. But we will follow-up the DMB to see if now this very much expanded information will also be made available, and we'll get that we'll get that to you. Moving on to Assemblymember Hoover.

Thank you. I think my question is for Renee. Just wanted to if you go into a little more detail for me on, how you feel like California's DMV regulation sort of compare to other states and, what you see as sort of the national landscape on these this this testing as well?

Thank you, Assemblymember Hoover. Yeah. So California is the only state that has this phase permitting process. It's the only state that has the mileage thresholds that I mentioned in my testimony. And also, the the data reporting is much more robust.

Specifically, only collision reports are required in other states and at the federal level. And here in California, you have new new requirements for system failures, immobilizations, the harsh braking requirement that I mentioned, and also now the notice of noncompliance that will take effect on July 1 and the vehicle miles traveled. So I think that's gonna provide the most transparency anywhere in the nation.

Alright. Did you have any others? Okay. Any other questions? I'll, ask a couple. Appreciate the question from my colleague, to you, miss Sanguinetti. I apologize for saying your name That's fine. Wrong at at the beginning. Because I do think it's important that you all have the information you need when there is an incident to be able to do that absent, you know that's under a company's control. Because I know that's the hardest thing when there is an issue getting information from the company is is hard having been a part of private industry.

And so if we're going to be collecting this information on behalf of the public, you know, I I support that it should be made available. To, the industry, a couple of questions. First, has it been clearly communicated to you all what steps your members would have to take in order to have a DMV action reversed once there's been this whole notice of noncompliance?

Thank you, madam chair, for the question. I I believe the answer is yes. I think that is spelled out in the regulations.

Clear enough for you all?

Yes. I haven't I haven't heard heard any concerns from my members to that Okay. On that specific point.

Thank you. And then, the incident in December, there were over a thousand reported AV vehicles coming to a stop to more than two minutes. And so the question is is that would that have in your understanding of the regulations, qualified as, something that had to be reported just on its own, not a not a, not a law enforcement giving some type of, notice of noncompliance or noting an issue, but just on its own, do you believe that, the company would have to have reported that?

Thank you, madam chair, for the question. So immobilizations have to be reported whenever a driverless AV is stopped in an active travel lane, cannot continue the driving task, and must be retrieved or removed by a human driver or a remote driver. And so I think it's up it's depends on the interpretation of the of the regulator and the enforcers, but that is I I think it's it's spelled out pretty clear in the regs at this point.

So when a vehicle is mobilized, not retrieved, but is immobilized for some period of time, that would not qualify according to the regs and your understanding of the regs.

Yeah. My under I would need to look back at the exact paragraph in the regs specifically, which I don't have in front of me right now, but we can we can get you a better answer.

Okay. Thank you. I just wondered from you all's point of view, not

Yeah. I mean, I think I think, generally, from our point of view, these the the new regulations were very clearly worked on for yeah. They're they're very in-depth. Okay. And and we have a good understanding of what they mean. Okay. There's also an explanatory statement that that was helpful as well that the DND published.

Good. Good. I'm glad I'm glad you feel that that way about those. Last but not least is so we know at the federal level, lots of advocacy is happening. And so I just wanted to understand, the position of advocating at the federal level to preempt state regulations of autonomous vehicle.

And is that as is that your all's standard, right, of of that you would prefer to see this done at the federal level versus the state?

No. Thank you very much for the question. And I and it's a little bit more complicated than a yes or no, but I am working on this at the federal level as well. So I'm happy to describe what we're doing and what our our position is. Our position is not that we should preempt every state AV law.

In fact, we go around actually asking for states to to pass laws so that we can operate Because a lot of states define that you have to have a human driver in the driver's seat with a driver's license. Right? And we don't have that once we get to level four autonomy.

And so we believe very strongly that states have their roles in terms of regulation, and that includes a lot of things that are within the DMV's purview and within the state's purview here, like licensing, insurance requirements, reporting requirements.

While we believe it is helpful for them to be federalized and standardized, we still think that states have their own role to play in terms of regulating AVs and that and what we are asking for at the federal level is preemption around design, construction, and performance standards, which is squarely under the purview of the National Highway Traffic Safety Administration.

Just to confirm, you said design, construction, and performance standards. And do you all feel that any part of the regulations that just came out from the DMV weigh into those three? Any one of those three?

So I I I am not a a NHTSA attorney, but From a general one. That would be a good question for NHTSA. But, generally, I think the DMV regulations stay within the boundaries of operational requirements as opposed to design, construction, and performance standards. And the reason we the reason those have to be federalized is so that you can have a vehicle cross state lines. Right? It's all about interstate commerce.

Yeah. Okay. Thank you. Looking to from the members, not seeing any others. Thank you so much. I appreciate, the discussion back and forth as well as your initial testimony.

Thank you, madam chair.

Thank you. Thank you.

Alright. And moving on to our second panel, I would like to now invite deputy chief Rabat with the San Francisco fire department as well as Allison Dretches and Rob Patrick with Waymo to the table. This is now an opportunity to speak about the requirements for first responder interactions, remote operations, and requirements for notices of noncompliance.

Alright. And moving on to our second panel, I would like to now invite deputy chief Rabat with the San Francisco fire department as well as Allison Dretches and Rob Patrick with Waymo to the table. This is now an opportunity to speak about the requirements for first responder interactions, remote operations, and requirements for notices of noncompliance.

All right. You may begin at your convenience, and there are five minutes for each group. Thank you.

All right. Thank you. Good afternoon, Madam Chair Wilson and members of the committee. My name is Patrick Rabbitt. I'm the Deputy Chief of Operations of the San Francisco Fire Department, and I'd like to thank you all for allowing me the opportunity to present this afternoon.

The San Francisco Fire Department, it's an all-hazards emergency response agency that protects the city through fire suppression, emergency medical services, rescue operations, hazardous materials response, and other disaster preparedness. While the majority of our calls we receive involve medical emergencies, we also respond to fires, traffic collisions, technical rescues, water emergencies, and major disasters.

One of our top priorities is ensuring that our personnel can respond quickly and safely to emergencies. We have been working closely with AV companies and welcome this technology, provided it is deployed in a manner that supports public safety and emergency response operations.

AVs have, on multiple occasions, interfered with the San Francisco Fire Department emergency operations both during active responses and while units are operating at the scene of an emergency incident. These interferences frequently require direct interaction between San Francisco Fire Department personnel and the autonomous vehicle, diverting firefighters and paramedics away from critical emergency response duties.

In such situations, SFFD personnel first attempt to direct the AV using the standardized traffic control hand signals. If hand signals are not relayed to the vehicle appropriately, the vehicle does not respond. Our firefighters and paramedics will attempt to establish communication with the vehicle's operating company through an exterior speaker or the microphone system.

Should the company representative be unable to remotely reposition the vehicle, control of the vehicle may be released to San Francisco Fire Department personnel for the purpose of relocating it away from the incident area. These interactions can and do delay emergency responses, essential emergency operations, and create additional safety concern at an emergency scene.

In many instances, concurrent direct interaction with the vehicle, our fire department personnel will request via our radio dispatch center to contact the autonomous vehicle operating company, direct them to reposition or remove the vehicle from the incident area. A proactive measure we take in our department is during emergency dispatch, we send Avoid the Areas, which is the geofencing that was previously mentioned in this-- today's committee.

Avoid the Areas, or ATAs, are sent for incidents where our emergency personnel are working in the street, occupying lanes of traffic for EMS or fire suppression responses. Currently, we send Avoid the Areas for 1,500 peat--1,500 feet, excuse me--for reports of a fire in the building, as well as ATAs for 1,000 feet for other calls such as vehicle collisions, vehicle fires, shootings and stabbings, other incidents where we'd have a lot of our personnel operating and shutting down the streets.

We send about a-- we send approximately 25 Avoid the Areas a day to the AV companies immediately with dispatch, and our department has around 500 calls daily for service. We would like to continue to work with the AV engineers to teach their vehicles our response patterns and build out notifications for their vehicles directly from our computer-aided dispatch system. Another issue that has come across our department's table recently in my new administration is the impact on sleeper calls.

Sleeper call's an internal term we've developed to describe a call where an incapacitated AV passenger generates a 911 emergency response. So on a regular basis in San Francisco, AV passengers fall asleep inside the autonomous vehicle. When the remote AV representative is unable to wake the passenger, a call is placed to our 911 dispatch center, and then a call is placed to the fire department to provide a response.

The fire department is then dispatched to Code 3 with lights and sirens to that location. In 2025, fire department emergency response was dispatched 284 times for passengers sleeping in AVs. In 277 of these instances or 98%, the passenger was woken up by fire department personnel and left the scene with no transport to a hospital.

In 2025, our units were cumulatively out of service for 113 hours, responding to sleeping AV passengers that needed to be woken up. In 2026 through May 28th, the fire department has already responded to 159 calls for sleeping passengers in autonomous vehicles. This year's data, 90% of these calls, the passenger woke up and declined further medical treatment or transport to a hospital. Total fire department time personnel spent on sleeper calls in 2026 is approximately 64 hours.

These calls do consume emergency resources that could otherwise be available for other medical emergencies or fire responses or other urgent incidents in our city and county, and we believe there's a good opportunity to work collaboratively with the AV companies to develop an alternative response, protocols that reduce unnecessary emergency dispatches while maintaining the passenger safety of the passenger in the AV. I'd like to thank you all for the opportunity to present today.

We're grateful to the state and this committee for your commitment to ensuring that AV companies, fire departments, EMS agencies, and all emergency personnel can continue to coexist safely.

Thank you. All right.

Good afternoon, Madam Chair and members of the committee. My name is Allison Drutchas, and I serve as Managing Counsel at Waymo. I also lead Waymo's autonomous vehicle compliance function, and outside of my work at Waymo, I've also served as an adjunct professor at UC Law, San Francisco teaching on emerging technologies and how to regulate them. We really appreciate the opportunity to provide our perspective on California DMV's updated autonomous vehicle regulations today.

So Waymo's been very responsive to California's regulatory and legislative processes, including being very active in the public comment process for these regulations, as well as the stakeholder process for AB 1777 authored by Assembly Member Ting, which established the framework for AV citations or noncompliance notices.

Waymo's experience over the past 16 years includes safely driving more than 90 million driverless miles in California alone and has proven that AVs are a vital tool to reduce unnecessary traffic injuries and fatalities in California. As regulators in states around the United States and countries around the world consider how to establish regulatory frameworks for autonomous vehicles, it's really true that all eyes are on California.

They really do view California as the first proven and comprehensive regulatory framework in the world. So consistent with these regulations, Waymo maintains a dedicated emergency response line, which is staffed 24/7 for first responders to call.

The regulations require that first responders and safety officials can reach a live operator within 30 seconds of making a request through the vehicle's communication device, and from past experience, including the December PG&E power outage in San Francisco, we understand the need for first responders and city officials to be able to reach Waymo in a timely manner. The regulations also introduced mandates regarding emergency geofencing, as was mentioned.

Upon receiving an emergency geofencing message from a public safety official, AV operators must immediately direct their fleets to leave or avoid a designated avoidance area within two minutes. Waymo informally developed a similar process that was mentioned here with San Francisco, which will now be the subject of very specific regulatory requirements.

And then finally, under the remote operations requirements, our remote support personnel maintain the immediate capability to immobilize an AV or facilitate manual override by a first responder, which was also mentioned. This means that first responders can decide how and when to move a vehicle when they decide that's appropriate. And with me today is Rob Patrick, who leads our Emergency Response and Public Safety Collaboration at Waymo.

Thank you, Allison, and good afternoon, Madam Chair and members of the committee. My name is Rob Patrick, and I lead Waymo's Public Safety Outreach program. I joined Waymo after 30 years with the California Highway Patrol for the specific purpose of working with the company and assisting us with working with first responders in the cities where we operate. While Allison has detailed a legal architecture on the renew regulations, my focus will be on how these requirements translate into real-world safety on California roadways.

At Waymo, we understand that regulations are only as effective as the operational partnerships we build on the ground. Our goal is to give first responders the tools they need to do their job and to design our vehicles and operations to fit their needs in emergency situations, not to rely on them to be AAA for AVs. To date, Waymo has trained more than 12,000 public safety professionals in California from fire, law enforcement, and emergency medical services.

This includes collaboration with California Highway Patrol and the California Office of Traffic Safety. And we've worked with first responders nationally. We've trained over 35,000 first responders. These are not one and done or one-way communication exercises. These are the foundation for extensive ongoing engagement, a two-way collaborative learning that constantly informs and improves our operational safety plans. Central to this effort is our First Responder Interaction Plan.

Waymo publishes and continuously maintains this document, providing public safety agencies with explicit instructions on how to safely approach, recognize, and interact with our vehicles. The plan specifies our dedicated toll free 877 first responder line, routes calls directly to our emergency response liaisons, and provides clear protocols for on-site support. The plan also addresses vehicle capabilities. Our vehicle features clear visual indicators inside the cabin so that an officer can immediately confirm whether autonomous mode is actively engaged.

If an emergency requires immediate physical intervention, our interaction plan details how first responders can access the vehicle and manually override the system to safely disengage the car. Finally, our automated driving system is engineered to positively detect and respond to active emergency vehicles. Whether via line-of-sight sensors or advanced audio perception, the Waymo Driver is designed to pull over, yield right away, and comply strictly with all traffic control directions from law enforcement.

We are committed to maintaining consistency across the industry and ensuring the technology remains a trusted partner to California's emergency services. Allison and I would be happy to take any questions you may have.

Thank you. With that, moving to members of the committee to see if there are any questions. Assembly Member Hoover.

Thank you. I think my question would be for Deputy Chief Rabbitt. Just wanted to get your perspective on how things have been going since the passage of AB 1777, the creation of this dedicated line between AV companies and first responders, and wanted to sort of see your thoughts from the public safety perspective.

Thank you for that question, Assemblyman Hoover. The passage of AB 77 and the first responder interactions with the AV companies have been great. We do have the First Responder Interaction Plan. I believe a few parties have spoke on that today. The numbers, our dispatch center will use those to call our fleet.

Our crews in the field don't have department-issued phones. We only have handheld radios where we communicate our request back to the dispatch center. That will be changing in the near future. But our dispatch center has had great success contacting them and having the AV situation handled. We've had some delays and we're still working on that with the AV companies as well.

Thank you so much. And then, I guess I would pass it as well to both of you or whoever wants to answer, but I guess, you know, what are your-- what is Waymo and--to the extent that you have knowledge--other companies doing to really ensure that, you know, there's better interaction going on between your company as well as first responders?

Since our earliest days-- thank you for the question, Assembly Member. Since our earliest days, we believed that collaboration with first responders was critical to our operations. And we have, since our earliest days in California, worked very closely with first responder in San Francisco and other cities as we've heard-- as you heard from me earlier. So I believe that is the key element is having those ongoing discussions, having the opportunity for that kind of collaboration.

In addition to what you hear-- what you're hearing about what's available to responders in the field, we have a first responder outreach team with more than 200 years of experience, and members of that team are available to first responders in all the communities where we operate anytime to resolve concerns, to answer questions, to ensure that we have that continuous line of communication.

Have you seen a reduction in any of the data of incidents involving emergency vehicles since the passage of 1777?

I can't speak specifically to the data--I don't have that in front of me--but I do know that, you know, as-- again, I'd reiterate that we collaborate consistently with first responders.

Appreciate it. Thanks.

Assembly Member Macedo.

Thank you, Madam Chair. Waymo has been operating with the general public for about five years in California, correct? What's your overall safety record in those five years?

Thank you so much for the question. You know, Waymo is pretty transparent about our safety record, and we publish a lot of data that uses industry best practices in terms of how to make comparisons to human driving, and in our first 170 million miles through-- of driverless operation through December of 2025, the Waymo Driver, as we call it, the Waymo automated driving system, was involved in 13 times fewer serious injury or worse collisions than human drivers.

So it's a 13x improvement over human drivers in the same areas where we operate, and at our current scale, that means preventing a serious injury crash every eight days. So that's a real positive impact on road safety.

And then my next question is, what's the difference between remote assistants and remote driving?

I can answer that as well. Some of the introductory statements got into this as well, but to reiterate, the regulations that we're discussing today do very clearly differentiate between remote driving, which is a defined term in 227.02, as well as remote assistants, which is a separately defined term. Remote assistants agents, by definition, do not perform what we call the dynamic driving task. So they don't drive the vehicle remotely.

Their job is to provide information and to make suggestions to the automated driving system to the vehicle. So for example, an autonomous vehicle might encounter a situation where it sees that a roadway is closed, might reach out to remote assistants to confirm, think this roadway is closed. Can you confirm? And the remote assistant would then respond and say, yes, it's closed, or no, it's not closed. So remote assistants agents do not continuously monitor a vehicle.

You don't continuously monitor a set of vehicles with some expectation that they're going to jump in and intervene. The way that it works instead is that the car--the autonomous vehicle--is driving itself and when it encounters something where it might want some extra information, might want a suggestion as to how to proceed, it's gonna reach out to remote assistants and then that specific question or request is then assigned to an individual agent to look at the situation and provide the feedback. Once they provide it, usually takes a few seconds.

They go back to a pool where they can be assigned to a different vehicle asking a different question. So it's quite different from a driving-- or a driver. Remote driving is where you would have driving happening. The actual-- the vehicle is actually being controlled from a remote location. Those are two separate concepts and they're addressed separately in the regulations in terms of the training, qualifications, and what's required for each of those is different because they're inherently different.

Thank you, Madam Chair.

Thank you. Assembly Member Hart.

The first time I'd heard about sleeper calls. That is a fascinating problem. Is there anything that Waymo is considering to address that?

Yes. Thank you for the question, Assembly Member. We're working very hard, as Chief Rabbitt will concur, with the San Francisco Police Department, other first responders to improve our processes. We never wanna unnecessarily burden first responder resources. We, of course, have to balance that with the need for our riders to get medical assistance when it's required.

So we're working hard to improve protocols with their suggestions and we'll continue to collaborate to ensure that we reduce the number of those calls to the degree we can possibly.

What is it you can do? I'm not even sure what we-- would you shake the vehicle or something?

Well, we just-- well, that's--

And to follow-up on that, because that was gonna be one of the things I asked, what are you doing prior to calling the fire department--

Excellent question.

--for those calls?

Yeah, that's an excellent question because, I think on the face of it, of course, it would appear that we have a sleeping passenger in the back of the car and we immediately call emergency services. Of course, that's not the case. We have a process we follow by calling the rider's phone, by calling into the car, by turning up the music. We have a process we use to attempt to wake or arouse the rider.

Do you change the temperature of the vehicle in terms of making it really cold?

I don't have all the specifics in front of me, but there are a multitude of things that we're trying and we're looking at some additional improvements based on some suggestions about certain decibel levels, certain specific types of sounds that might more readily awaken that person. So we're working through that now, but we are collaborating with them.

And, again, it's that balance between what might be someone who appropriately elected to use our service late at night and may have fallen asleep in the back of the car and someone who used our service who may, in fact, need immediate medical attention.

I wish I could sleep that well, but is--

Hydraulics might be in order. I think this is a perfect time to go old school and bring back the hydraulics.

Is there anything else when you were collaborating, you know, sounds effectively that, you know, you've identified that needs tweaking in the regulations that we are, you know, trying to implement in California that you see around the corner? And, like, what-- five years from now, what would success look like in collaboration?

Oh, well, it's-- I mean--

How do you figure out a problem that you've been talking about?

Assembly Member, it's a fairly broad quest-- you know, broad concept, but I think success looks like clear communication and collaboration. I think that's what success looks like. When Chief Rabbitt feels like if they have a concern, they have someone they can reach out to who will sit down at the table and have that discussion and come up with an improvement for the process. In my opinion, that's what's successful.

Okay. And Chief Rabbitt, do you feel like you're on that path?

Yeah. Thank you for the question, Assemblyman. Yes, we do. With Waymo's team, we've had tremendous collaboration. Our agency has been kinda at the forefront of this with the City and County of San Francisco having a large deployment of autonomous vehicles, whether they're in testing mode or the fully autonomous with passenger mode.

And we had some really good personnel come up with an idea to start tracking the AV incident. So that's why we started gathering data related to AV stalling out and-- or coming into an emergency scene, or that we came up with the sleeper term because our forms kept showing they weren't in the scene but they were generating an incident. So then we came up with another metric to track and we've been working collaboratively with Waymo and the other AV companies as these issues come up in attempts to address them.

Thank you.

Thank you. Assembly Member Papan.

So, thank you all for being here. At San Mateo County, we certainly have a lot of presence of Waymo in my district. Very quickly though, how does first responders-- how are they able to access if we do-- on behalf of the tired population? How are you able to access? Is it that Waymo remotely unlocks the vehicle so that you can get in?

We could speak through the AVs. They're all a little bit different, just like all of us who have vehicles probably have a different make, model of a vehicle, but the autonomous vehicles do have a remote microphone and we could get the attention of the remote operators through that microphone and speak with them directly if we're at an incident with a AV that's experiencing some kind of trouble.

Okay. Just curious. Well, here's to getting more sleep.

Yeah. Thank you. I will say, like Assembly Member Hart said, that whole sleeper car, I had a whole different thought when you said sleeper car and then when you explained it. And that is a lot of man hours wasted, and so glad that that is-- that there is some sense of collaboration to help resolve that and figuring out what to do.

Be very much interested if you could provide to this committee afterwards the exact protocol for how you-- what you're doing before the call is made. I would appreciate that greatly. This-- I have a couple questions for Deputy Chief Rabbitt and then following up for Waymo. So for Deputy Chief Rabbitt, you noted that it was your desire to build out notifications directly from dispatch. Can you spend a little bit more time on that?

Yes. Currently, I did bring to this committee meeting our chief at our dispatch center, and he has been serving there for six years as this AV trend just really took off throughout the state, and what they did at our dispatch center for the police, fire, EMS calls, and what we're gonna be operating in the street, generated an Avoid the Area, which is geofencing, as we have been discussing here, and we chose those numbers in collaboration internally and brought it to the AV companies that we've been in communication with just based on the average block or street size in San Francisco. It's typically 500 feet.

So for the vehicle accidents, the auto pet accidents, shootings, stabbings, we'd have significant resource, but we thought 1,000 feet would adequately divert the AVs. If we instantly set it when we send the police, fire, EMS response, that would give them time to clear and avoid that area. Larger for a significant disaster; a fire or collapse, we set it up to be a larger Avoid the Area at 1,500 feet.

So the moment the dispatch is getting a call about a stabbing, or about a shooting or something, they're immediately--

Yes.

--right away? Okay.

It's in the AV company.

Thank you. That is very helpful. Secondly, you have a lot of Waymos. Waymos are built like a typical vehicle, meaning that if you had to get in and do a manual override, you're going to be able to see a steering wheel, you're gonna see a gas pedal, a brake, all those good things.

However, there are other companies making headway in San Francisco, like Zooxs, who don't have those things. And so, can you tell me a little bit about how you're managing an override or how you've been considering the override of that vehicle in comparison to a traditional AV-- well, I guess, a traditional regular vehicle, I guess, we're moving to. That's the traditional vehicle, but you get what I mean.

Yes. Thank you for the question. Regarding the other AV companies and their design models, we do continue to follow the First Responder Interaction Plan. We do have tremendous collaboration with Zoox as well. The manual override of a vehicle without a steering wheel, an accelerator, or a brake, we have not addressed yet, but we are still in communications with Zoox as they are expanding their services in our city as well. But I have no definitive answer for you right now, and I could get that for you, Madam Chair, if you'd like later.

Thank you. I appreciate that. And then to Waymo, this goes back to a conversation we've been having around licensure, licenses, driver's license. And so, over half of your remote assistants are located overseas. Remote assistants are required to have a license, and so what steps do you all ensure that they don't have points or things on their license? And then as it relates to your remote operators or drivers, which have to have a license as well, the same question in regard to, like, the point system.

Yeah. So we have, you know, confirmed that any of our agents who are--regardless of where they're located, they have to have a valid driver's license. In terms of international presence outside of the United States, we specifically chose places where we knew that we could verify that they're-- had good driver's licenses and their driver records. That was part of the selection criteria. So we've also monitor their driver's licenses like you would in-- with a United States-based license.

So we do that for both. And with respect to remote assistants, as we mentioned, you know, there is not a requirement that they are physically present in the United States or licensed in the United States. With respect to remote drivers, you know, if we were to have a remote driver, they would be located in the United States and would have a United States license.

Sorry. So there are no remote drivers which operate the vehicle based on the new definition, or defined definition-- there's no remote drivers living in other countries? They all-- there's--

That's correct. There would-- there are none.

But I thought--and correct me if I'm wrong--I thought they could have a-- except commercial driver's license because a commercial driver's license is a different beast and you all are-- were talking about your, you know, standard vehicle. So if you have a-- when you hire a remote driver, they have to have a U.S. driver's license, not a foreign driver's license? I thought they could have a foreign driver's license.

I would love to get back to you. I think that's not something that we-- when I said they'd be in the United States, that's Waymo's position.

Okay.

But, yeah, I think there are requirements for monitoring their driver records and things that, you know, may or may not be challenging if they weren't in the United States, but, yeah. We'd have to confirm what the parameters are there.

But what is consistent is that you all, in hiring remote drivers, no matter where they are--we're not sure where they are yet--but no matter where they are, you all look to ensure that you can verify their driver's record so you make sure they come from countries where that is something that you could do as a part of your hiring practice?

That's right. For all of our remote assistants, yes.

Okay. And then, as we saw in December, there was a whole issue with communications, but just in general, because it could be because there's a blackout, but it could be for other reasons, you know, a cyber attack, whatever, so what redundancies do you all have in place to ensure that communications with remote drivers, operators, assistants, whoever is-- if it's severed, like, what redundancy you have in place in case it's severed?

So I think it depends on which aspect of this we're talking about. So in terms of supporting the vehicle, our Waymo vehicles are capable of operating for some period of time without connectivity in order to achieve what we would call minimal risk condition, come to a safe stop. If during that short period of time where they're still able to drive, if connectivity is then restored, they could just continue driving without any interruption.

In the event of-- you know, for whatever reason there's a longer term connectivity issue, cars are designed to come to a safe state and come to a stop and then be recovered at some point later by-- we have a roadside assistance team. We also partner with tow operators in all the places where we operate.

So that would be kind of what happens to the vehicle. As far as reaching Waymo, that's where the number would come into play, so the toll free number could be a way to reach us if you see-- if a first responder, for example, were to encounter a stopped vehicle trying to reach out, and connectivity from the vehicle itself was not available, that's where the phone number could be useful, or we have QR codes as well that they can use to reach out. So that would be kind of the best way, I would say, to reach out in that context in addition to, you know, the vehicle being in a stop safe position.

Okay. All right. I do not have any other questions and I don't see-- oh, sorry. Vice Chair Davies.

Thank you. A question to the deputy. Have you ever had the experience when the remote driver actually got disconnected to the car?

Thank you for the question, Vice Chair. As I understand this, if our communication was discontinued with the remote operator, is that how-- we have not had any documented incidents of where we're communicating and we lost comms with the remote operator of an AV yet.

And then for the other two, do the remote operators, do they have to be drug tested, and is this something that's continuous? Thank you.

It depends on the-- you know, which role we're talking about. With respect to remote assistants, we do, at Waymo, do require certain drug and alcohol testing, including preemployment and random and poor-cause sort of checks. So it really depends on the role in terms of what the requirements are, but the regulations do address the requirements for each type of role: remote assistants, remote driving for light vehicles, remote driving for heavy-duty vehicles, et cetera.

Alright. Oh, Assembly Member Carrillo.

Thank you, madam chair. What is your projection as to when these autonomous vehicles will be available for consumers for purchase?

I think we we don't have a date, for any of that. I I will say that it's Waymo's ambition to build the Waymo driver, and to deploy our lifesaving technology in in many different forms. But unfortunately, no specific dates for private ownership at this time.

No preorders yet?

Not yet.

Alright. I don't see any other questions. Thank you so much to our panelist. Now we'll move into our, final, panel, where we'll discuss the regulations related to heavy duty autonomous vehicles. We're gonna first start with professor Scott Mora with the Institute of Transportation Studies who'll be testifying remotely.

As that is set up and he begins to speak, I'd like to invite Matt Broad representing the Teamsters and Chris Irmsen, with Aurora to our table. And at this section, each of our speakers have five minutes. Okay. We can't hear you, so hold one second. I don't know if that's him or you.

We'll see. Okay. Okay. Professor Mora, can you hear us? If we can't hear you, though, just keep talking as we figure it out.

You can, sing your ABCs for us. Okay. Let's see. Is there anything I need to do up here? Okay.

No. No. They're they're they're figuring it out. Our technical folks are working on it remotely, and so we will wait patiently while they do that. This would be a great time for members of this committee.

If you haven't got some refreshments, some water, or taken a quick comfort break, I would recommend to go do that right now. And from the members of the public. You all as well.

Test one two three one two three.

There we go. There we go. Okay. We're gonna wait a couple minutes for people to return. So if you're under the sound of my voice, now would be a great time to return.

You have sixty seconds. I said two minutes, but, really, I only meant one. Alrighty. Professor Mora, take it away. You have five minutes.

Okay. Madam chair Wilson and, committee members, thank you for having me today. Thanks for your flexibility to participate remotely as I help my parents move into their house that was burnt down in the Altadena fires about a year ago. So I'm a professor at the University of California, Berkeley. Hopefully, you can see my slides.

We can.

Okay. Great. So I'm a professor of civil environmental engineering. I'm also the acting direct director for Berkeley's Institute of Transportation Studies. I'm also the director for a DriveAI consortium that brings together industry and California agent government agencies together around autonomous systems.

I am also the co principal investigator for a US DOT, national center focused on autonomous freight. First, I I just wanna let, you know, my own perspective, which is that as a university of professor, hence slides, we're really at ITS Berkeley and UCITS. Our mission is to drive innovation that shapes transportation in California and beyond through leading edge research and active collaboration with policymakers and practitioners. So I deeply appreciate this opportunity to, participate in the conversation and learn from all the other panelists. Okay.

So this panel is focused on heavy duty autonomous vehicles. And I first wanna start out with why do heavy duty autonomous vehicles matter. And the first one I'd like to make is that freight movement is absolutely essential to California's economy. I'm talking about ports, logistics, agriculture, manufacturing, interstate commerce. Obviously, California has, the the largest ports in North America with a Port Of LA and Long Beach just as one example.

The other important point that I wanna mention about these new regulations that allow heavy duty autonomous vehicles is that large trucks represent a small fraction of registered vehicles, but are involved in a meaningful share of row of fatal roadway crashes nationally. So their profile in terms of, crashes and injury inducing crashes is different than light duty vehicles, robo taxis, and the like. And we need to think carefully about that.

And that can be simply understood by the following fact is, heavy duty vehicles generally have much they have greater weight by definition. When they operate at higher speeds, they're they're they're they have more kinetic energy.

So then they have much longer stopping distances, which means that if there is a crash, there's greater consequences. This is why we have drivers that are highly trained human drivers that are highly trained in how to operate these types of machinery in in the event, of of an incident. And and the last point and and perhaps even the most important is what this could mean in terms of improving freight efficiency, but also safety, which are crucial public policy objectives.

What I always like to say is that, you know, autonomous vehicles make sense if they accomplish three things. They improve safety on US roadways, number one.

Number two, they improve safety on US roadways. Number three, they improve safety on US roadways. If autonomous heavy duty vehicles also have the benefit of improving our freight efficiency, we can think of the supply chain crisis, you know, during during the COVID era, then then that could be an added benefit too. But we have to think about what it does to economic development, how it distributes, you know, economic development within California's ecosystem. Now what are the new regulations accomplished?

So for the first time, California now has a pathway for regulated deployment of autonomous heavy duty commercial vehicles. And so this simple graphic just at a super high level shows that, you know, AV manufacturers will put together safety case along with other, piece of information that'll go to the department motor vehicles for review. They then can do testing, graduate eventually to deployment, and there's ongoing oversight with the different reporting requirements. The key features as highlighted earlier were the safety case review, I think very, very critical.

Operational design domain restrictions, so defining how and when, these vehicles can can operate.

The, permitting requirements are now much more, granular in, what the requirements are in terms of, miles they they need to drive. For example, heavy duty vehicles will need to accumulate half a

million miles before they can graduate to the proceeding level in those three three levels. There's ongoing reporting requirements as as we've heard, and then the new regulations really provide enforcement authorities both on DMV side and some action for notice of noncompliance by by peace officers. So so those are positive directions for these new regulations. But, really, I wanna highlight that the real test of a regulatory framework is not whether it allows or prevents deployment. It's whether it generates the evidence needed to advance California's policy goals.

So so that's where I'm really focused is how are these regulations generating evidence to determine if we're advancing policy goals? And and I'm focused on safety, and we can also talk about free efficiency economic development. Now there's a shift towards more informative safety metrics that we've already talked about before. There was really just three categories, disengagements, crashes, and vehicle miles traveled.

You know, now, you know, after the new regulations were implemented, we we now have this concept of dynamic driving task failures, vehicle mobilizations, heartbreaking events, and then, of course, crashes and vehicle miles traveled remain.

Importantly for crashes now, when you're in the deployment stage, you need to report crashes as well. Okay. This is the important piece. The context for heavy duty AVs with respect to these new safety metrics. The first thing I wanna mention is regarding immobilizations.

If we're talking about heavy duty vehicles like long haul freight on highways, when there's immobilization on highway travel lanes, that can be exceedingly dangerous. Right? There there there isn't always a shoulder where they can pull over safely. So it it's unlike a robo taxi in a city urban environment where you could pull over to the side to the curb and and and potentially be out of the way from the flow of traffic.

In a highway setting, there's much higher risk, so so we need to consider that more carefully.

Heavy duty AVs, as I mentioned before, have much longer stopping distances. They also accumulate vehicle miles traveled much faster. So so we can't compare the safety of vehicles operating in a higher speed freeway environment the same way that we do in an urban environment. The denominator, is just different for that different use case, so we need to be careful when we assess safety. The other thing is that while most highway driving with heavy duty vehicles is relatively boring, it's straight at constant speed.

When there is an an incident that could cause a crash, they tend to be more rare but more consequential. And, I think one of the previous panels, at least saint sanguieti, mentioned this point. So so the risk profile for heavy duty vehicles is different. The other point I wanna make is the need for leading, not lagging indicators. That is some of these indicators identify an issue after the consequence has occurred.

We need metrics that really identify issues before a high impact consequence can can occur. And so some of these metrics, dynamic driving tasks, failures, vehicle immobilizations are towards that, so I see that as a positive step forward. And as was mentioned earlier, I also wanna highlight this, is data should be public to allow peer reviewed research. So previously, with all the data on disengagements crashes, it allowed a lot of peer reviewed research with the even the limited data that was available to assess safety.

And so you can, you know, allow the research community to help assess if that data is open.

A a a couple more points. Learning from other states. So heavy duty vehicles, commercial vehicles already operate in other states. Five, to my knowledge, Texas is the strongest example. But the California context is, of course, different because we have these dense urban freight quarters.

I think of 07/10 where I grew up in Southern California, going from the ports up into the Inland empire. You're you're going through an urban environment there. Now while other states are generating operational experience and and we have a panelist here who who who who can probably talk about that from from their company's experience. You know, California, I think, is generating regulatory experience. And how do state agencies work with these private companies to do this in a way that's that's as safe as possible?

So I'm gonna leave you with, I think, four questions here, Right? The California should continue to evaluate safety. And I think it was doctor Corey Jackson who asked the question at the very top, which I think is the question that needs to be asked over and over again. It is, are, in this case, autonomous trucks improving safety performance on roadways? Are we reducing those forty thousand deaths on US roads every year?

Data. So the data that's being collected, the question to ask is, are the reporting requirements producing meaningful insights that we can assess these policy goals. The third is scalability. So when and under what conditions should operational restrictions evolve or be loosen as they improve? And and I think there's a a lot of positive movement in the regulations to provide much more granularity to move forward or backward, depending on what the case might be for the appropriate AV OEM.

And the fourth one I'm just gonna leave you with as a food for thought, maybe you ask question about it, is, you know, California also has zero emission vehicle goals. So how do these come hand in hand? Right? Can heavy duty AVs accelerate heavy duty zero emission vehicles? There's some intersections there that I think are worth discussion, but I'll stop there.

Thank you.

Thank you so much. And and and people probably felt like that was longer than five minutes. I did give more time because it is a framing conversation, and I thought it was beneficial. And so we're going to move on to our other two panelists, and you may begin at your convenience. I don't remember the order now.

I think you're first, Matt.

Okay. Thank you, madam chair and members. My name is Matt Broad, and I'm here on behalf of Teamsters California representing 250,000 workers across the state. This is probably unsurprising to you all, but I have a less rosy perspective on these regulations and believe the Department of Motor Vehicles has taken California down a dangerous path. On April 28, the DMV formally adopted regulations opening California's public roads to the testing and deployment of heavy duty autonomous trucks.

These rules remove the long standing prohibition on autonomous vehicles weighing over 10,000 pounds, vehicles that can gross up to 80,000 pounds on our highways. This means for the first time ever, trucking companies are gearing up to put self driving big rigs on California's roads. Teamsters California believes these regulations are fundamentally defective legally, analytically, and as a matter of public policy. We raised these objections throughout the rulemaking process. They were not adequately addressed.

Let me explain why. To start, the DMV's economic analysis is fatally flawed and attempts to skirt California's administrative procedure act.

Under the APA, an agency must assess the costs and benefits to businesses and individuals in California. And if those costs or benefits exceed $50,000,000 in the first year, the agency must conduct a detailed standardized regulatory impact assessment known as a SIRIA. Here, the department claimed that the deployment of heavy duty autonomous vehicles would result in zero job loss for California workers. We challenged that at every step in written comments at hearings and before the office of administrative law. It's not a serious position.

Heavy duty trucking employs hundreds of thousands of Californians, and the explicit purpose of this technology is to replace them. The notion that removing drivers from commercial trucks at Scale Produces zero employment impact defies logic, it defies economic literature, and it defies the lived experience of workers who have watched automation hollow out industry after industry. The DMV also ignored the most basic implementation implementation costs. Standing up these operations in California, training law enforcement, first responders, the remote operator technology, and so on and so forth.

But the flawed economics are only part of the problem.

The safety framework is equally deficient, relying on self certification rather than independent validation. The DMV accepts a safety case prepared by the manufacturer itself with no independent audit, no external verification, no external verification, no enforceable performance threshold. We are asked to trust that companies will police themselves when the stakes are human lives. That's not how we regulate aviation. It's not how we regulate pharmaceuticals and it should not be how we regulate 80,000 pound vehicles on California's public roads.

These regulations compound this problem by keeping the public in the dark. This legislature passed AB 3061 legislation Teamsters California cosponsored to ensure Californians could at least see the data, including mandatory public reporting on all collisions, citations, unplanned stops, and disengagements in open machine readable format with real financial penalties for violations and the ability for the public to submit incident reports directly. The governor vetoed this bill. What the DMV adopted in its place is closed, manufactured to to regulator reporting structure.

The public cannot see the data.

The legislature cannot easily access the data, and there is no independent fine schedule to enforce compliance, nor do the regulations establish meaningful operation limits tied to California's actual conditions. Companies can accumulate the vast majority of their testing miles in Arizona or Texas and then deploy here in our mountain passes, our fog, our wildfire smoke, our dense urban corridors. That's not a safety standard. It's a loophole that puts speed to market ahead of public safety. And that loophole did not appear by accident.

These, regulations were rushed through with little opportunity for public transit transit or, excuse me, public challenge. The DMV held only one public workshop before proposing them. The comment periods were narrow and short, and the April 2025 draft to the final adopted text, the d the department consistently retreated from stronger safety requirements, every change running in the same direction towards whatever industry wanted. The consequences of getting this wrong are not abstract. Monthly crashes reports reached a peak of a 110 incidents in May 2025.

The DMV has received nearly a thousand collision reports, and that's before a single heavy duty autonomous vehicle has been commercially deployed. We've already seen AVs block emergency responders driving to crime scenes, make illegal u turns during DUI enforcement, and now apparently, our fire firefighters are are babysitters for drunk passengers. That's a new one for me. Now imagine those failures at 80,000 pounds at highway speed. The DMV has made its choice.

Now this legislature must make its own, and we cannot stand by while heavy duty autonomous vehicles are deployed under conditions that endanger our communities and eliminate good paying jobs with no independent safety validation, no honest economic accounting of economic harm, and a rule making process that prioritizes market signals of republic safety. If the DMV will not put people first, this legislature must, by enacting strong requirements for the safe testing and deployment of autonomous vehicles and protection of middle class jobs. Thank you.

Thank you. Alright. To our next.

Chair Wilson, Vice Chair Davies, members of the committee, thank you for the opportunity to be here. I'm Chris Ermerson. I'm the cofounder and CEO of Aurora. Over two decades ago, my team at Carnegie Mellon University took part in the DARPA Grand Challenge. We were in the desert trying to prove that we get a truck to drive a 100 miles across sand and rock with no human at the wheel.

I took part because I thought it'd be fun and interesting. As my work progressed, the massive opportunity of this technology became clear. Forty thousand people are killed on US roads every year mostly due to human error. That's not acceptable, and we must do better. When I think about Aurora's technology, I think about my own children.

I remember the day they started driving. I want them to share the road with drivers that never get tired, that never look at a phone, that never have a lapse in judgment. This is why I cofounded Aurora in 2017. I believe then, as I do now, that deploying automated driving at commercial scale is the most important contribution I could make to society. Since Aurora's founding, we've had deep roots in California.

Our employees live and work here. I've been proud to call the state home for the last seventeen years. When we think about where this technology should operate, California is not an afterthought. It's central to what we're building. We're here not just as a company, but as Californians working to do things the right way.

In April 2025, Aurora became the first company in the world to commercially launch autonomous trucks without safety drivers. This launch was the result of a careful and thoughtful process. Our safety case relies on millions of pieces of evidence, and we've had independent experts review our safety processes. We also engage with federal and state regulators including regular briefings for CalSTA, the California DMV, CHP, and members of the legislature. Importantly, we did not launch until we knew that our trucks were going to be safe on public roads.

Since then, we've driven over 370,000 miles. Over those miles, we've seen that the safety, sustainability, and economic benefits are real. Even so, I know there are questions about the impact this technology will have on Californians who drive trucks for a living. I want to address those directly. First and foremost, we should appreciate and thank anyone who is willing to drive a truck.

The trucking industry is the backbone of our economy, and drivers into your grueling schedules, face some of the highest injury rates of any profession, and suffer on the job fatality rates that are 10 times higher than the average American. When I talk to leaders of trucking companies, they always lead with the importance of safety and follow that with how essential drivers are to their business and how they see automated driving complementing, not replacing the people who drive their trucks.

They see AV technology enabling a better quality of life for their people, their employees. Our recent partnership with McLean is a perfect example. Their president talked about how autonomous trucks will allow their drivers to focus on the critical last mile remaining the face of the company to their customers.

I also believe that autonomous trucks are not coming to displace workers. They will be deployed on the routes that are least desirable, most dangerous, and most demanding. I'm confident that as we introduce automated trucks, we will see the role of our logistics professionals elevated. And as the Aurora Driver rolls out, drivers will enable to focus on short haul routes where they are sales reps, customer service agents, and the face of their company. And importantly, they'll be able to get home tonight to their families.

Automated trucks are not coming overnight. They will integrate into logistics businesses over years. I'm here today because as California implements this new pathway to autonomous trucks, we want to share our real world experience having engaged in the DMV's years long regulatory process. I welcome the opportunity to discuss this and the gradual integration what the gradual integration will look like. I first participated in DMV's workshop on automated vehicles almost fifteen years ago when AV legislation was first introduced.

When the DMV shifted to focus on autonomous trucks about five years ago, we will continue to engage with the DMV's public information sessions, spoke at workshops, and participated in several rounds of comments. I appreciate the deliberative and thoughtful process that the department has taken to ensure that California voices have been heard. And while I believe the resulting regulations are onerous in some regards, they provide the clarity and certainty that enables us to continue to build and invest in California.

Aurora's technology will make logistics safer, more efficient, and more sustainable, and we'll work through the regulatory process to bring these benefits to California. Most importantly, we look forward to building the future deliberately, transparently, and shoulder to shoulder with partners in industry, labor, and government.

Thank you for the opportunity to be here today and look forward to answering any questions.

Thank you to the panel for your remarks. Now moving on to members for comments starting with assembly member doctor Jackson, and then I see you over there, Erins.

I've first, before my question, our majority leader had to leave, so I just wanted to read some comments that she wanted to make sure was on the record. And so our our majority leader wanted to thank UCITS for always being a resource on this and other issues, the Teamsters for their years of collaboration, and Aurora for taking the time to explain all of their work on safety over the years.

I am concerned that we will see heavy duty autonomous vehicles in our neighborhoods and on our city streets because of how these regulations are drafted. These vehicles can still go on streets with a speed limit of two 25 miles per hour or less if the truck is traveling to a shipping, distribution, fueling, charging, or hub facility. This leaves a lot of uncertainty for communities.

I am also worried that autonomous trucks would be traveling on high speed rural roads that do not have a shoulder or as many safety markers as a freeway might. Can you just share with the committee more about how the conditions on the roads and risk factors will be accounted for in the permit permitting process?

Sure. Happy to. Thank you for the question. And sorry she can't be here, but we'll go from here. So first and foremost, I think the the DMV did a very nice job of steering the regulations to push trucks to be on the places where they need to go.

It turns out across The United States, 90% of distribution centers and warehouses are within two to five miles of freeways and interstates, and that's the place we're going to go. That's where it is useful to haul goods between these points, because our trucks can't do the local goods delivery part of this. As I mentioned, the role of the driver in those operations is very different. It's not just driving the vehicle, but it's representing the company, engaging with the customer, and all of that.

So it's a difficult thing to express, I think, but the the intent was clear, and it's our intent to abide by that.

So we're excited for that. As it comes to rural roads, in the near term, we don't expect to be operating on rural roads. We tend to be offering on the central corridors of freight much like we do in in other states, whether it's in the dense urban cores of Dallas and San Antonio and Austin, or whether ultimately we get the opportunity to do that in California as we go through the regulatory process.

On rural roads, so I do think there's a unique advantage to our farm communities. One of the advantages that automatic vehicles have is they will increase the range that produce can reach while maintaining its freshness and thus extending the shelf life for it.

Or alternatively, expanding the reach of which you can sell into as a farmer or distributor without having a reduction in shelf life. So I actually think that will ultimately be a huge advantage, but it's not what we're working towards today.

You know, one of the only things that I've obviously, the question that I continue to harp on is the idea of what is a safer way of delivering goods. Is it with human drivers? Is it with automation? Is there a mix of both? You have Aurora has also operated in other states or have been going through the testing phase in other states.

What is your what what does data tell you? And I think at some point, madam chair, we really probably just need an objective data to really answer the question once enough data is available in terms of what produces the best safety and prevents the most fatalities. Because I think, for me at least, that's a central question because we also need to make sure that we are balancing between technology, workforce, and safety at the same time.

And I think that we're at some point we need to get to that point. But what what is your your thoughts in terms of safety issues?

And Teamsters are more than welcome to chime in as well. Yeah.

And I think professor Mora may have thoughts on it.

Professor as well.

I'm I'm happy to start. Yeah. I think it's critical. You know, safety is underpinning our ability to operate and deliver, and it's core to what we do. Our company's mission is deliver the benefits of self driving technology safely, quickly, and broadly.

And those words are in that order for a very good reason. We have to put safety first. If it's important, we wanna move with urgency to do it. And then, ultimately, we want this benefit as many people as possible. So as we think about decision making with the company, it is with that paradigm.

When it comes to deploying our technology, we use the safety case process, which is very similar to UL 4,600, quite aligned with it. And we end up having millions of points of evidence that we look at to get to conviction that it's safe to put this vehicle on public roads. We've also run studies to understand the safety implications of this directly. So we did a study looking at fatal accidents on I 45 between 2018 and 2022 that involved heavy trucks.

I think there was 24 of them.

And we took the police reports, we recreated the accidents in our simulation system, and then asked what would have happened if the Aurora driver was driving the initiating vehicle, effectively the vehicle that caused the crash. And the answer is none of them. Every one of them would have been avoided. And not one of them was due to the driver making some miraculous Mario Andretti esque swerving maneuver. It was really because the driver has the superhuman ability to pay attention consistently and constantly.

One of the ones that sticks with me is two light vehicles were had a fender dinner on the side of the road. The people got out to exchange information. The truck came along. The operator of that truck saw them saw the collision late, avoided onto the shoulder, and drove through the people. And that's two families that will never be whole.

It's tragedy. In contrast, when we ran the simulation, what happened with the Aurora driver slowed down, moved over two lanes, and everybody would have gone home. And that's an impact that you just can't understate. We did a similar experiment looking at the I 5 Corridor near Sacramento over the same time period, twenty eight to '22. There was, I think, 15 or 18 collisions that involved passed or involved heavy trucks that led to fatalities.

And once again, we had the same outcome that because of its vigilance, the Aurora Driver is able to avoid this. And it is equipped with capabilities that people can't have. Our first light lidar allows us to see further beyond headlights than a person can see.

The fact that we have 360 degree vision means that the Aurora Driver can see all around it, and not just see all around it where I would have to turn my neck if I were attempting to see there, but can actually pay attention to all of it. And that means that it can do things that a person can't to keep the public safe on the road and, of course, to to keep itself safe.

So while we don't yet have the statistical evidence for this, the analysis and efforts that we put in our safety case process, and then the anecdotal evidence we're beginning to see on the performance in real world conditions is pretty profound.

Thank you. Through the chair. I would say safety is a huge priority for us. Right? We're a labor union.

Go ahead.

We represent workers that have to be on the road, whether they're driving in their personal or professional capacity. And we we share that commitment and, you know, would argue there are lots of ways to continue making the road safer, without necessarily needing to remove a human safety operator from the vehicle. Of course, there are infrastructure changes. There are also hours of service restrictions that we fought for that have been turned over by various administrations.

And so I think there's a lot of work always to be done on safety, and and we're committed to that.

I I would actually, commend mister Irmsen and and welcome the comments that he made about believing that human, humans have a a place within the, within the driving, world, particularly in the short term capacity. That's great to hear. I have a hard time reconciling it with the fact that our majority leader has a bill, AB 33, that requires a human safety operator for local commercial deliveries, and his company is consistently opposed to those bills.

And so actions speak a little louder than words to me, and we would welcome continued partnership with them if that's something they wanna talk to us about, about actually making sure that workers are protected. And then if the technology can't do local driving because it's not there yet for the reasons that he specified, that we're making sure that the human safety operator is there for the safe deployment.

So I would I would say that. I would also welcome sharing public data and see if the companies are are willing to commit to that. I think there's a lot left to be desired about the data reporting, a lot of, frankly, spectacular claims about what simulations have shown and not shown. And I'm not trying to disparage those.

I'm just gonna say that in real world driving conditions, when you talk about things like mud spat splattering on lidar and no one there to clean it, How does that impact the technology?

These are things that we're not gonna see, frankly, until this technology is actually deployed. And in the process of testing and deploying it, why would a company oppose having a human safety operator as an added layer of security and safety, until we can actually verify the claims that are being made by these companies?

Thank you for that. I'm not sure if, professor, if you had any thing you like to add?

Just just a quick comment on your question, doctor Jackson, of our AV SAFR. Let me tell you what the peer reviewed literature says. In short, it's yes and no. Let me unpack that slightly. So the peer reviewed literature in the last couple years has shown some evidence that some of the AV companies that are focused on, you know, light duty passenger operation, you know, robo taxis have been able to achieve a lower rate of injury related crashes, some of the AV companies.

Nevertheless, the no part of the answer is that there have been stress points. And we've already talked about, interactions with first responders blocking ambulances, the mobilizations that can, change the flow of traffic. So so we clearly see potential, but some some growing pains. So so that's my answer to your question. Yes and no.

Of course, you would, professor. That's exactly what I expected. I think the last thing I'm sorry to take up too much time, madam chair. I just wanted to make sure obviously, I represent part of the Inland Empire where trucks are on a daily for us and nightly. Emissions is always a big issue for us.

How do you believe your technology would be able to improve the efficiency of supply chain? And most importantly, for my district, how would ensure that we actually have less emissions as we're as goods are being transported.

Yeah. This is this is one of the really insightful points to make here, I think. So thank you. Automated vehicles have an immense number of opportunities to improve the quality of life for us. So, obviously, safety is critical and core to this and is kind of the the foundation that we build off of.

But the sustainability benefits are real. So today, at our fleet, we see a 14% improvement in fuel efficiency relative to the 2023 kind of national fleet average of fuel consumption. That's profound. Right? That's the equivalent of one in eight trucks off the road in terms of emissions.

And so that's great for our customers because that reduce the fuel cost, but it's also great for the environment in terms of emissions. And so when I think about us being able to deploy this broadly, do that safely, and have that kind of impact, it's really, kind of incredible.

Thank you, madam chair.

Thank you. Moving on to Assemblymember Arons.

Thank you, madam chair. Listening to both of you, it we have, like, a tale of two cities here about what the future is going to be. And I hear both of these points of views in my town halls in Silicon Valley where a lot of the innovation is created for the world. And I guess my question would be, you know, human drivers experience over years behind the wheel.

I certainly didn't add wouldn't admit it at the time, but when I started driving at 18, I definitely got into a lot more accidents.

And I guess my question is, how does Aurora use these sort of simulations, and how can you more specifically speak to the concern, you know, that Teamsters, that my constituents or others are bringing up about deploying this technology? Because I I really do think that there are major differences of opinion here. But I just wanted you to maybe more pointedly speak to some of the the criticisms of this technology.

And I would love, Matt, if you could sort of help respond more pointedly as well, because I do think, even to the professor's point, it could be, you know, and and, but I really would like to get more pointed answers from both of you in terms of how this technology is being used, how it has improved in the past five, eight years. And and and, honestly, what concerns would you have that maybe that, you know, where are we missing the mark?

And and we talk a lot in the legislature about in a bipartisan basis about, you know, the legislature not exerting our authority enough or giving it away too much and how the legislative branch is not as strong willed as it perhaps it should be, especially and not even ever being able to really anticipate the how quickly this technology has moved forward. At the same time, we don't wanna stifle that innovation. And, madam chair, if I can just combine questions for the sake of time Yes.

My my last question would really be is I'm worried about workforce. I'm worried about augmentation versus replacement is what we often talk about with innovation and technology and and and safety.

And I'm wondering if the both of you could speak to how prepared our workforce can be in the future. I know one point of view is that it's going to lead to job loss as a lot of innovation and technology and advancements has led to. Another point of view is that it creates workforce. It creates jobs.

I'm wondering if you both can speak to that more pointedly, because that is a probably on top of mind when when I'm thinking about how we adapt this technology is is what is the future of work?

What is the what is the future of how you know, more pointedly, what types of jobs will this create? Are we prepared for that? Is our education system prepared? Are community colleges or other institutions prepared for the these jobs of the future? And and how do we grapple with this?

Because the legislature is grappling with these issues. Silicon Valley is grappling with moving forward. We certainly want you know, my primary concern is keeping jobs in the state of California, making sure that our tech jobs stay in the state of California, but making sure that, you know, not everyone from Silicon Valley works in tech. And so how are we how are we keeping people employed for the future? And is there a way to adopt this technology and still keep good paying jobs?

Because that's the bottom line of, I think, what why what what is most concerning to me is, you know, how do we not stifle innovation, but how do we not really lead to that job loss?

So I I think these are obviously important questions to wrestle with. So maybe I'll start there there was a triple whammy of questions. So maybe I'll I'll start with the first one. I know you can do it. I I will work through it.

So the first, you talked about the the role of the DMV and the regulations and the strength of the assembly. I guess, there's some sense that maybe the DMV did a decent job since I'm not particularly happy. He doesn't seem particularly happy. So maybe they got to a decent point of of balancing things. So like I said, what what we value is the fact that there is clarity that we can go work with.

The things we don't love about it, but we can work with it. On the human driving experience simulation, and how do we know that that things are working well. So maybe take a step back and talk a little bit about how we get to the conviction that we have a safe product. So for us, that involves a safety case. And that safety case has got five core elements to it.

The fact that the vehicle has to be proficient, it has to be fail safe, it has to be continuously improving, that we have to be resilient. And then as a company, we need to be trustworthy. And it's important that they look at this not just as a technology, but as a integrated holistic view of how we build the thing from the people and how we operate as a company to how the technology works on the road.

When I'll just maybe just touch in on the first part of this, which is the performance element of it. And that's where we run millions of tests before we ever release something. And the way we come up with those tests is is multifold. So one is just driving around, you see a lot of normal stuff that happens. And so we can turn that into test very easily. It doesn't take a whole lot of thought to to get that. We can get a really a whole lot of novel effort to get data.

The second is when something interesting happens on the road. And so, for example, we had an event where we were down in Houston, Texas. The truck approached a traffic light. The light was green. As we began to approach the intersection, the Aurora driver saw a vehicle coming from under and underpass to our right, accelerate through the intersection, crash into the vehicle to our right.

The Aurora driver had seen it coming and stopped and avoided that collision. So that was great outcome in that moment. But the question is, what would happen if the Aurora driver is a little bit sooner or a little bit later? If it's going a little bit faster or a little bit slower? So we'll take those kind of rare events that we see on the road and generate variations around them to get us conviction that that kind of family of interesting events have now been tested.

And then finally, there's the things that we never see. Right? The fatal accidents are comparatively rare on our roadway even though there's a lot of them. And so there, we lean into the NHTSA crash taxonomy, which describes all the way different vehicles can crash. Then we run through a process kind of diabolically thinking about the bad ways this can happen, and generate those simulations, and ask the question, what would the Aurora driver do faced with that?

This incredibly rare, incredibly challenged situation, and ensure that the behavior it takes in those situations is appropriate and safe. So that's, like, kind of how this works. So how do we see that it actually translates into practice? So we had, one of our customers brought their ambassador drivers, which are some of their most experienced drivers to come visit our trucks and see our team and ride along with them. Experienced drivers to come visit our trucks and see our team and ride along with them.

Those drivers came in much like many of us would be skeptical. I think one of the drivers said something about how, you know, how could this possibly go wrong? I've seen this movie, maybe it's the Terminator, something like that. He spent a day with our team, and importantly, they went out for a road test with the truck. And the story that came back from the day was night and day.

One One of the women who was one of these experienced drivers came back amazed that it used the turn signal, and that it would signal appropriately. And she just wished all other drivers would do that. Another gentleman came back and said, you know what I do? I just put my truck in cruise control and follow it because I know it's going to be safe, and I'll be safe behind it.

And the gentleman who started the day talking about this, you know, the Terminator, how could this go wrong, came back saying this was Optimus Prime, who for those of you who are not up on your, you know, transformers lore was the leader of the good guys, big truck that transformed into a robot.

So that was pretty rewarding to hear folks who really are deeply experienced in the industry and spend time there. Look at what we've built and say, yes, This is this is something I would like to share the road with. In fact, I'd be happier with this on the road. And then the the last question you talked about was around workforce development and navigation. So I'll reiterate what I said in my cost.

We should be willing like, we should thank anybody who was willing to drive a truck. It is a very difficult, dangerous job, and we need it. We need people willing to do that. At the same time, we should be thinking about how can we elevate their role, how can we create roles that allow them to work in logistics, but enable them to be home at night with their families and have safer and better life outcomes? And that is what we think we're doing.

And we're not just hoping. We're actually investing to make that happen. So we have cross trained drivers into different roles at Aurora to build that experience. We've worked with Gallatin College in Montana to develop training for the next generation of LIDAR technicians. We've worked with Over the Road Garage in Texas to develop training for, automated vehicle truck technicians and mechanics.

And we recently committed a million dollars for workforce development in a program we call Aurora Works to work with other institutions to create training programs to help elevate, again, the logistics professionals into other roles in this industry, and there will be new roles. I think the best metaphor to think about in terms of the job and opportunity here is the inventor of the automated teller. And when the automated teller came was first deployed, the fear was that we would have no longer bank jobs.

And what happened was the automated teller made it easier to open open branches. It create more opportunity for employment, and it moved the teller from a job that was purely a cost of taking a check and counting out dollars into something that actually contributed to the profitability of the bank.

And thus, there's more tellers today than there were before the automated teller was invented. So I have a lot of optimism and belief in how this will actually positively employ implement impact the workforce.

Thank you. You know, I just reiterate that the jury is still out on the efficacy of the technology. Don't take my word for that. Take the professor's word. There's lots of literature out there, particularly when you look at miles driven by humans every year in this country versus what's been done by autonomous vehicles.

There have been several high profile instances. I believe, Aurora may have been one of them where, there was, driverless technology, and then one of the client companies, I think it was PACCAR, requested that, that drivers go back into the vehicle. And so, again, we're looking at a very short amount of time to kind of make these decisions. At the same time, I wanna touch on intangibles. Right?

Because since we've seen the deployment of this technology, we're learning about the ways that they may operate longer, right, or be able to not drive drunk, but they also can't close doors when passengers leave them over open for Waymo. They can't wake up drunk passengers. They can't put emergency cones out if they have to pull over for the side of the road on the highway. There are a million things. The Waymo incident with a thousand vehicles that went out.

We're not gonna be able to sort of anticipate these novel events until they happen, which speaks to sort of the need for, as policymakers, you should be wanting to do the most to, okay, we're gonna explore where this technology go goes and its promise, then we ought to do it in a way that maximally protects safety and jobs. And and, you know, on that point, I would just say that we have to be real about what the technology is about.

It's partially about safety, but it's about eliminating labor costs. That's the selling point. Eliminating labor costs for long haul. We believe also eliminating labor costs for short haul. We have, like I said, a piece of legislation by a by a committee member that this committee has seen that that requires a human safety operator for that short term that we're talking about.

And so this is the opportunity for the legislature to step up and not just take mister Irmsen's word for it, but actually create a a statutory precedent that allows for the deployment of this technology in a safe way, but also allows for the deployment in a way that's protective of the jobs that he claims he wants to see protected.

Well, the debate continues then, mister chair. Thank you.

Thank you, Assemblymember. Next is Assemblymember Hoover.

Thank you, mister chair. So I think, you know, you've to mister Irmsen, You have sort of dove in already to the simulation and explaining those. I think my curious my question would be, you know, I think we have a lot of data on crashes that are caused by human factors. Right? Of distraction, fatigue, impairment.

What what is the data that you have seen in other states as as I believe Aurora is testing more extensively in other states? Have you seen an actual reduction in the data, similar to what Waymo has presented, for example, where you can actually compare, autonomous vehicle miles to, you know, directly to human driven miles and and see an improvement?

Yeah. What what I can say is that today, the Aurora Driver has no at fault collisions to its record. It's been out there operating driverlessly for some time. The events that we have had have been during testing, and there's been situations where other people have hit us on the roadway, and the Aurora drivers work to avoid those collisions even then.

So I did have one question for, the professor. You know, I I appreciate the studies that you presented on, the dab that we have currently on whether these trucks are safer. I think my question for you would be, will they be safer? Or maybe a better way to frame that is do they have the potential to be safer? Because to me, as policymakers, I really think that is the more important question.

I think this technology is obviously at a very infant stage even today, And so there's certainly not going to be conclusive data, yet that these trucks are safer on the roadways today. But I think the question is, you know, do they do you believe they have the potential to be safer, and what data points would you be looking for to to determine if if that is indeed the case?

Yeah. Thank you, member Hoover, for the question. So so just to be clear, the the peer reviewed studies that I mentioned were not focused on heavy duty vehicles, not heavy duty vehicles. They're focused on light duty vehicles, you know, robo taxis and and the like where there's some evidence from some, but not all, automated vehicle OEMs that they are safer with respect to injury related crashes and in particular with interactions with vulnerable road users. That's like pedestrians, bicyclists, and the like.

Okay. Your question is about heavy duty vehicles. Do they have the potential to be safer? I think my personal opinion is that they have the potential to be safer. But there are some there there will be some some growing pains and some challenges.

Right? Let me, again, emphasize that the way I see it is that the risk profile by the way, when we're talking about heavy duty vehicles, we're mostly I'm mostly talking about long haul freight on highways. We can also talk about heavy duty vehicles that are in ports, drayage, mining, you know, those things. So let me put those aside and focus on long haul freight, on on highways. The risk profile looks different because crashes are even more rare, but they tend to be more severe.

I think one issue is that if we're gonna measure safety by injury related crashes and we just have to wait for them to happen and enough of them to happen to actually assess if it's safer, then that might not be something the public is willing to absorb. Right? Willing to experience, willing to put up with. That's why we really need not lagging indicators that evaluate safety after the fact, but leading indicators.

So I think DMV did a pretty reasonable job with this when they had this idea of dynamic driving task failures, heartbreaking, things like that, things that are precursors towards safety.

I think there is more that could be done, honestly, in terms of metrics. And and the last point I'll I'll make, is that I think it's incredibly important that this data that's reported to to DMV be made public. Because once you make it public and open, then you enable a worldwide community to look at all that data and assess from a third party point of view. Is this safer?

With the type of data that was being collected before these new regulations, it was hard to say a lot that was definitive. I hope I answered your question.

No. I appreciate that. I think there's a lot of valuable information in that answer as we continue discussing this topic moving forward. I think, lastly, I guess, I'll I'll ask Matt a question just to close this out. You know, this this example of the bank teller, you know, what we obviously don't know what the future holds when it comes to workforce, when it comes to potential job opportunities.

I will say that while there has been in the past technologies that have displaced jobs, I think that's undeniable. There in in many cases, those technologies actually didn't fully replace workforce. They actually expanded economic opportunity, maybe grew the economic pie, for example, and created new jobs that that, you know, workers could go and do. What what do you see in response to that bank teller example? Do you see a world where that is possible?

And if not, you know, why is, the replacement of jobs most likely outcome versus, you know, maybe an increase in in workforce needs?

Sure. I think, the ATM example is one example, and we could probably find others, like the mechanical loom that eliminated weaving entirely. Right? And and so it really depends. It's fact specific.

I will say that autonomous vehicles are a form of AI. And in this moment, you know, everybody's constituencies, everybody you talk to is concerned about how AI is different with respect to mass displacement. And I think, it will be, potentially unique and in a way that will require government intervention to, to hold companies to the prod the the, promises they're making. Frankly, if this is a technology that results in more driving jobs that are better, I think that would be something that we would welcome. We are skeptical of that.

And without the sort of government intervention that steps in and says, hey. You know, we need to think about mindfully where technology makes sense, where it doesn't, where there are unique safety concerns, then, we're we're just sort of leaving it up to a type of technology that is specifically geared towards job elimination, not just displacement elimination, and trusting that everything will work out fine. And, of course, we have serious reservations about that.

Thank you. I think I I would just, sort of end with you know, I think the hard part about this as policymakers, right, is we don't exactly know the full outcome here. I do understand there being a concern over workforce. However, I also truly believe that in the long run, these these, technologies will absolutely make our roads safer, mostly because, you know, mostly because, you know, I think that there are no matter what, we know the data of human drivers.

We know the data when it comes to crashes caused by humans, and and it is certainly a statistic that we would love to see reduced substantially. So I think with that, I appreciate the dialogue today, and thanks for for coming to this hearing.

Thank you, Assemblymember Hoover. Assemblymember Lackey.

Yeah. Thank you. I don't have a lot of questions. I just have a couple remarks that I I believe are worthy of expressing. First of all, I think as representatives of the public, our first obligation is to people.

And I think sometimes considerations blur that and allow us to look elsewhere. Because and I think what what needs to be considered here is that we need to make technology a partner, and we don't necessarily need to surrender to technology. But I I do have frustrations, because sometimes we get this dichotomy of opinion. It's an either or, and I I think we have opportunity for synthesis here. But we have circumstances as we adapt.

I think we should always go with the most safe, and that that includes all added layers of assurance. And I I don't understand why there's such a resistance to go with the most safe while we're transitioning to these heavy vehicles adding another level of insurance or assurance. Why is that opposed to? Why why is there such heavily resistance? But we've seen it.

As we know, we've seen legislation before that tries to incorporate a hybrid. And I I I think that that is usually the best pathway forward. And to assume that any technology is a 100% reliable is a fallacy. I mean, we're talking about technology and the beauty, and there are some beauties to technology. No doubt about it.

But we saw in this hearing today, the unreliability of technology. We couldn't communicate with one or one person or another. It's not a 100% reliable. There's no such thing. Right?

And so we need to temper reason with what we're doing. And I I think that driverless circumstances are gonna be a prominent part of the future. There's no question about that. What I'm asking for is cooperation. Let's let's do this together to where we don't alienate one group or another. We we can actually do this in a in a reasonable way to where everybody wins. And that's what I'm hoping for.

Thank you, members. Assembly member Lackey. I wanna thank all the panelists today that does not just the three. Oh, Assembly member of CITO. Sorry about that.

Thank you, mister chair. You are speaking my love language. If we can help ag, one thing that we are different than most industries is as soon as we remove that crop from the ground or we have that product that we've gotta get it to the end consumer as quickly as possible because it wants to spoil. So I liked what you said about speeding up this food supply chain. But beside that, you know, we're dealing right now with lots of bottlenecks in the food supply chain. So how can your technology, help agriculture specifically?

Yep. I think that's that's exactly it. Right? Is that we can help move goods more quickly, while maintaining, if not increasing safety. And then they're available whenever.

So if the harvest is coming in late in the evening, the vehicle can be there ready to go. If it's first thing in the morning, it can be there ready to go. You know, we're the the kind of the level of flexibility that will come along with that, I think, is pretty profound.

And the other thing that I just Kinda wanted to elaborate on is it it if I'm hearing claims that there's some people that are claiming the whole process, through the DMV and the legislature and the agencies felt rushed, but it seems pretty extensive from what I've seen. So can you just give me, like, a CliffNotes version of, a, why you think that it's been adequate that we've had this oversight, and then, b, just what you've been through to get your product to market?

Yeah. So thank you. I think that's a great question. And I think it's really important to not lose sight of the fact that this whole process started in fifteen years ago, when the legislature, invested in the DMV the opportunity to put in place regulations for automated vehicles. Like, at that time, perhaps even appropriately, the DMV said, we'll focus on light vehicles.

There's nobody working on heavy vehicles today, but we will come back to heavy vehicles because that was the intent of the legislature. Like I said, that was fifteen years ago. Aurora has been working on automated trucks for eight years. And along that path, you know, the company's been around for almost ten years now. We regularly engage with CalSTA, CHP, DMV, and talked about the importance of this, the importance of having this technology serving Californians.

It took another three years, before the Department of Motor Vehicles was able to seriously engage on the topic. Then it's been a five year conversation, where all stakeholders, industry, safety advocates, labor have been involved in part of that conversation. It's gone through, as far as I can tell from the outside, multiple internal rulemaking revisions that ultimately led to public common process. So it as a company, we have been forced to go invest in other places rather than California because there's been no certainty or clarity.

Now that we have it, we may not love all elements of the regulations, but we know what we have to do. And when I think about the opportunity for safety, I think we're all on the same side of the table. But we wanna see our roads safer, you know, that our our colleagues in labor wanna see the roads safer. And, you know, as a policymaker, I assume you all wanna see the road safer. So we're excited to finally have made it through that extensive process.

Well, as a legislator that commutes from the Central Valley to come up here to work, I I want the roads to be safer. I I see lots of distracted driving happening. I see the horrible accidents that have happened on the road. So thank you for your dedication and your resilience because I it sounds like this has been a long path.

It sounds like you have an open door policy that you want to engage, so I hope you come true to that, that we kinda all need to come to the table.

This technology is coming, and California cannot be left behind when we should be the leaders on all things supply chain. And I think this is just another tool in the toolbox. I do not think this is a replacement for the wonderful men and women that work in the supply chain and otherwise. So I wish you the best of luck, and thank you for being here today.

Thank you, Assemblymember Macedo. And, again, thank you to all the panelists throughout the day. It's been very informative. Appreciate chair Wilson organizing this informational hearing and all the team putting together the panelists. It's been really helpful and and useful.

Now we're gonna open up comment from the public. Please just line up here and, limit your comments to as briefly as you can. The hour's getting late. And one minute would be ideal. We don't have the microphone just yet.

I was I've been cutting sentences out of

We're we're adaptable. But yeah. Thank you.

Hi. Mark Watts representing Contra Costa Transportation Authority tonight. They they're very pleased to to see that the regulations have been developed as a true collaborative. And they have long relied on their Comentum Station facility, AV testing facility in Concord to better understand the world of connected and autonomous transportation through, pilot programs. In particular, CCTA is encouraged by the regulations as they apply to AV transit and large vehicle applications in the regulation.

Many transit agencies face challenges including operational costs and others, but the management of CCTACs that AV tech technologies can help expand citizen access to transit and improve transportation safety and better serve a range of communities. And I'll conclude with that. I have much more to go, but thank you very much.

That was perfect. That was sixty seconds. Yes. Thank you.

Good evening. Charles Watson on behalf of Zoox. As we've as we've heard today, the DMV has recently concluded a very robust public process, to modify the rules governing the operation of autonomous vehicles on California roads. This included the new requirements established by AB 1777 and passed in 2024, And we would just ask, and strongly urge that the legislature allow these regulations sufficient time to take effect before considering new legislation to to make changes. Thank you.

Thank you.

Thank you, mister chair. Jack Yance, behalf of Consumer Technology Association. CTA represents more than 1,200 member companies, 80% of which are small businesses and start up membership spends the autonomous vehicle ecosystem, including suppliers, manufacturers, software developers, and service providers. CTA supports a clear policy framework that enables the safe deployment of autonomous vehicle technology in California. California has long been a leader in innovation.

The state should avoid creating new barriers that could slow AV deployment, testing, or development. AV policy should be tech tech neutral, performance based, and focused on measurable safety outcomes. CTA encouraged legislation to work closely with industry, regulators, first responders, and other stakeholders to ensure any future AV framework supports safety while preserving California's leadership in transportation innovation. Thank you.

Thank you.

Good evening. Sophia Quach on behalf of the Bay Area Council. The Bay Area Council is proud to represent members of the AV industry that are working to bring safer, more efficient, and expanded mobility options to Californians throughout the state. We urge the legislature to allow the DMV's newly finalized AV regulatory framework to play out and avoid layering further legislation on top of this framework.

Burdening a massive multiyear regulatory overhaul with legislation before it even has a chance to take effect threatens California's economic competitiveness, stifles, tech leadership, and disrupts a robust public safety structure.

The dubious regulatory process has been a multiyear effort of stakeholder engagement, draft revisions, and public workshops involving labor, local governments, and industry experts, as well as law enforcement. The updated framework creates an incredibly high bar for operation with some of the strictest standards in the nation. We ask this committee to support stability.

Let's allow the DM freeze framework to take effect on July 1 and measure its success with data before adding additional regulations on a technology that is already meeting the highest regulatory hurdles in California history. Thank you.

Thank you.

Good evening, chair and members. My name is Stephanie Jimenez, and I'm here on behalf of Chamber of Progress, a tech industry association that supports policies to ensure everyone benefits from technological progress. Our partners include transportation innovators like Waymo, Zoox, and Aurora, but they do not have a vote or a veto over our positions. We all share the same goal, making sure autonomous vehicles operate safely on California roads.

The question today isn't whether oversight is needed, it's whether the legislature should impose new statutory mandates just weeks after the DMV completed the most comprehensive update to its AV regulations in nearly a decade.

On 04/28/2026, the DMV finalized major new rules. These regulations strengthen oversight and expand enforcement tools, require detailed reporting of immobilizations and safety events, set new standards for first responder communication, and establish clear rules for remote operators and assistance. We discourage any redundant or conflicting regulations that lock rigid requirements into statute, such as those regarding staffing, response times, and vehicle design that could quickly fall out of step as a type as a technology evolves.

The DMV already has expertise and authority to adjust its requirements based on real world performance, stakeholder input, and new safety data. That flexibility is one of California's greatest strengths as a leader in AV innovation.

Before adding another layer of prescriptive mandates, we should allow these new regulations to take effect, evaluate their performance, and let the expert agency do its job. Thank you.

Thank you. Good afternoon, chair and members. Ashanti Smith on behalf of the Silicon Valley Leadership Group representing California's leading technology employers, including several autonomous vehicle companies operating in the state today. Just wanna quickly frame that Silicon Valley didn't just adopt this technology. It birthed it.

The companies defining autonomous vehicles worldwide were founded and built here, and it's not just the companies. It's the engineers, the capital, the suppliers, the research universities, and the entire ecosystem that exists nowhere else on earth. Our ask is a stable statewide framework that is led by experts already doing this work at the DMV and the CPUC that ensures safety, reliability, and allows for existing regulations to take hold. Predictability keeps the ecosystem, the investment, and the jobs here in California. Thank you.

Thank you very much.

Hello. Matt Lege with SEIU California. We like to acknowledge the work of the DMV and their process, but we do feel like there are steps that, additionally could be taken, by the legislature. For example, we think that there needs to be significantly more public, display as talked about in the hearing of information data so that we have some idea of what's going on, where the crashes are. We think that, there, are still some critical gaps regarding public safety and interaction with law enforcement.

For example, there's no response time. There's no actual deployment of a personnel to deal with someone like sleepers like they were talking about today. These are some real world issues that continue to arise that we think that the legislature is well within their purview to, take on. So thank you very much and look forward to continued conversation.

Thank you for the comments.

Thank you, mister chair. Louis Costa with Smart Transportation Division. So we represent bus operators and school bus operators. They're currently engaging on the same roads with robo taxis like Waymo's and Zukes and others. Daily concerns, safety concerns, daily problems, with, you know, short stops.

And all of a sudden, you've got an articulated bus that has to deal with that. You have a school bus that has its sign out, flashing lights. Robotaxis are running around there, causing safety concerns for the kids and for the for the passengers of the buses. So this the new regulations don't prevent big rig trucks from sharing those same roads. It's disingenuous to say you're just gonna go from point a to point b from Arizona to to a warehouse out in the desert somewhere.

They're gonna be coming. That's what they want. They wanna take those all all of those short hauls and long hauls, and it's gonna cost jobs. We would urge the legislator to require a human safety operator on board during any testing or any deployment until we can make sure that the data does say that it is safer. The professor said it wasn't necessarily safer.

It could be safer, but we don't know if it's gonna be safer. And we urge you to to take the safe course and make sure that we have human safety operators on board.

Thank you.

Thank you, mister Krasta.

Good afternoon. My name is My name is Liz Fishback. I serve as the director of state and local affairs for Stack AV. Stack AV is a Pittsburgh, Pennsylvania based autonomous trucking solutions company. We upfit Peterbilt five seventy nine's class eight trucks with level four technology, which allows, as you've heard today, the truck to drive itself in certain locations within the operational design domain, whether that's meteorological, topographical, time of day, those sorts of things. I know there was a slide that showed that five states have activity.

It's far, far more than that. We personally test in Arizona, New Mexico, Texas, Louisiana, Alabama, Florida, Georgia, West Virginia, Pennsylvania, Ohio, and are expanding even further. That's north, south, east, west, red, blue. We choose these states, like our friends and other companies, because they have clear and consistent paths for both testing and commercial deployment. We are thrilled that after fifteen long years, the DMV has finally released those regs.

We look forward to working with the DMV to safely bring this technology to California roads and your constituencies to help bring safety to public roads and efficiency to the logistics supply chain. I also do wanna point out that the DMV regulatory process does have a step involved in which a driver is already required in that process. So thank you very much for your time and the length of the hearing today.

Thank you for your comments. Good day. The last word.

The last word. I will keep it brief just because it's been a long day for y'all. John Kenrick on behalf of the California Chamber of Commerce. Viewing this d v DMV regulatory package is a bridge between today's public concerns and the future of autonomous transportation. The rigor that it has can help build public confidence while it's clear and responsible pathway enables continued innovation, investment, economic development, and job creation in California.

Strong safety standards and technological progress should reinforce one another, and we believe the regulations advance both. Thank you.

Thank you, John. And we are adjourned.