Assembly Standing Committee on Privacy and Consumer Protection

Good morning, everybody. I'm going to call the Assembly Privacy Consumer Protection Committee. So if we can all take our seats, that would be fabulous. Thank you. Hello. I did. I was gonna say, most of the people in the room don't miss me very much because we were together 12 hours ago. Actually, wait less than 12 hours ago.

So welcome back. I hope you all slept in the intervening hours. Welcome to the Assembly Consumer Protection and Privacy Committee. We have 16 bills on the agenda today. Just for everyone paying attention. File item 8, SB238, Smallwood-Cuevas, has been pulled by the author. The bills on consent are SB278 by Senator Cabaldin and SB504 by Senator Laird.

Will take a motion on the consent calendar. When we have a quorum. To effectively manage our time, we'll be limiting testimony to two witnesses for support and two witnesses in opposition on each Bill. Each witness will get two minutes to present their testimony.

And after they conclude, everyone in the room can tell us whether they like the Bill or dislike the Bill. And we at all times, allow for testimony, written testimony, through the portal on the website. And we will, if it is okay with Senator Wiener, begin as a Subcommitee, because he is here.

Thank you, Senator, for being so prompt. So the first Bill we will take up is SB53. Senator Wiener, when you're ready.

Thank you very much, Madam Chair. And thank you to you and your staff for working with us. And I accept the Committee's amendments as outlined in the analysis. Last year, as you recall, because the Bill went through this Committee, I authored comprehensive AI safety legislation, Senate Bill 1047.

The Bill passed with bipartisan support in both houses of Legislature, and the Governor did veto it. But in his veto message, the Governor was very, very clear that the risks that we had identified in SB 1047 were real, that it was important for the state to take action to try to get ahead of those risks.

And even though he did not agree with the particular approach in 1047, he made those statements and then convened as part of his veto message, a working group with some leading minds in the space to come back with recommendations.

That working group, which included one of the most Vocal opponents of SB 1047 as well, came back with its final report about a month ago. And we had been advancing SB53 with whistleblower protections in the creation of Cal Compute, the public cloud for the bulk of this year.

And as I had indicated earlier in the year, we took a look at the working group report, and there were a number of very solid policy recommendations in that report. And so we then incorporated them into SB53.

So, you know, I'm very proud, and no offense to any colleagues, but I'm very proud to represent the beating heart of AI innovation in San Francisco. We have a powerful and growing ecosystem of AI startups. And so much amazing innovation is happening in our great city, but in other places as well, with this important and powerful technology.

And AI has the potential to make the world a better place and to improve life for humanity, to help us cure diseases, to help us tackle problems that we thought we weren't sure if we could tackle, whether climate change or so many other issues. And it is really exciting what AI promises.

But as with any powerful technology, in addition to benefits, there are risks, and that's normal. And it's not about eliminating risks. Life is about risk. Unless you're going to go live in a basement somewhere, and even then the ceiling could collapse on you.

It's about trying to understand the risks and then trying to get ahead of them in order to reduce the risks. And I know through this Committee, various bills have moved through around deepfakes, around algorithmic discrimination and various kinds of risks.

The risks that we're focused on in SB53 are more catastrophic risks, whether around chemical, biological, radiological, nuclear weapons, whether around major cyber crimes, threats to critical infrastructure, and so forth. And so what this Bill does, and SB 1047 was a different approach having to do with requiring safety testing and liability.

This Bill is not about liability, it's about transparency. And so it will require transparency for large, large companies. Safety and security protocols. Our goal in this Bill is not to sweep up startups.

And so that's why we have to qualify a large 10 to 26 flop as the size threshold and per amendment that we worked out with the Committee requiring that the company also have at least $100 million in annual revenue. So they'll be required to be transparent about their safety and security protocols and risk evaluations.

They'll be able to redact those to protect trade secrets.

The Bill will also require that when there is a critical safety incident, for example, like a loss of model control or model enabled you, biological weapon threats, major cyber attacks that requiring the developer to report to the authorities within 24 hours and to file a report with the Attorney General within 15 days.

The Bill grants the Attorney General some flexibility to update thresholds, as this is an evolving technology and the only enforcement lies with the Attorney General for penalties for violations of the act and not reported, not transparency. There's no new liability for harms. This is a transparency Bill, not a liability Bill.

So in addition to the whistleblower protections and the creation of Cal Compute, the public cloud to enable researchers, startups to be able to have access to increasingly expensive compute.

And by the way, we've already gotten positive signals from some folks in UC about, not on behalf of UC but on behalf of themselves about the possibility of hosting Cal Compute at UC, which would be amazing. This Bill is very, very solid.

I do want to say that we, because the Bill went into print with the working group recommendations incorporated just a few weeks ago and now we're, now we're seeing significant amendments that we've accepted from this Committee.

We are the engagement with stakeholders, including the industry, including the Administration and obviously the Administration, given the veto, we need, you know, make sure that the Administration is going to be on board with what we're doing. That engagement has begun and it's going to accelerate.

And I'm going to continue to have a very open door and we are proactively soliciting input from large companies, from startups, from researchers, academics, investors. Anyone who has feedback, we want to know that feedback. Anyone who has ideas, we want to know those ideas.

This Bill, you know, we're going to continue to have that engagement and have an open door and of course we'll stay very close with the Committee as we move forward and as that feedback comes in.

And I think this is a very solid Bill in the state that it will be amended if it passes out of Committee today.

And I believe, and my goal is that for to produce a Bill ultimately that we can send to the Governor that we can be proud of and the Governor can be proud of, I hope signing so with that I respectfully ask for your aye vote.

And with me today to testify is Steve Newman who has co founded eight tech startups including Writely, which is now known as Google Docs, and Terry Oley with Economic Security California.

Thank you. Two minutes each when you're ready.

Okay, thank you Madam Chair and Members of the Committee for this opportunity to testify today. So, yeah, my name is Steve Newman. I'm a software entrepreneur. I've sold companies to Google, Apple, Sentinel 1. I'm now President of the Golden Gate Institute for AI. But here in a personal capacity.

I learned to code at age 8, which means I've been in tech since 1975. I played my own small part. We built Google Docs in the attic over my garage. But this is the most exciting generative AI is the most exciting technology I've seen.

I've also never seen any technology less well understood by the people who are building it. Any AI developer will tell you that what we're doing here is not is as much alchemy as engineering. Silicon Valley was built on innovation and innovation is built on information. But we're lacking information today.

Generative AI is so powerful, evolving so rapidly and frankly so mysterious that we can't predict when catastrophic risks will occur. There are warning signs. OpenAI's own preparedness framework cites systems that will soon have the capability to create meaningful risk of severe harm. That's why Governor Newsom's policy working group called for immediate action.

This bill's reporting requirements and whistleblower protections will provide the information we need to proceed with confidence. And I know a good technical compromise when I see it. That doesn't mean both sides are unhappy. It means you've put in the work to get what you need at minimal cost.

As you've just heard from Senator Wiener, the working group and the authors of this Bill have done that hard work. We should take advantage of this opportunity to safely advance this once in a generation technology. For these reasons, I urge your support of this critical legislation. Thank you.

Thank you, Madam Chair and Members of the Committee. I'm Teri Olle with Economic Security California Action, proud co sponsor of the Bill. And of course I want to thank the Senator for his continued leadership on this issue.

I'm going to talk for a minute about Cal Compute, which would, as you know, establish a consortium to develop a publicly owned and operated cloud computing cluster so that we can expand and democratize access to this critical AI infrastructure. Cal Compute first addresses the problem of extreme concentration of compute power in a few hands.

Compute is the most important factor in enabling or limiting AI innovation. It's extremely expensive and controlled by a tiny number of companies who have an outsized say in what's developed and whom it benefits. Meanwhile, startups and even major academic institutions lag behind. For example, Stanford only has 300 of these specialized chips needed for AI.

Microsoft has 1.8 million. Disparity affects career pathways too. With California educated tech talent lured away from solving our pressing public problems because the private sector is frankly more dynamic, this is a significant loss.

Second, Cal Compute will continue California's legacy of transformative public investment in cutting edge research institutions such as the pioneering Stem Cell Research center and our world class national labs. We are the home of AI and we should be leading on this too. New York is already underway and showing what's possible after investing in a consortium last year.

Empire AI has 200 plus researchers busy building climate models and improving cancer diagnoses. California, join with them and show the world how to spur innovation in a climate of transparency and safety. And finally, Cal Compute is popular, as we've heard. It received nearly universal support as part of last year's 1047.

And since then it was named as the top recommendation in the Little Hoover Commission's report on AI, which cited reasons similar to those I offered above, among others, including recognizing the benefit of the State of California achieving a degree of technological Independence, which seems more important than ever. For all of these reasons, I urge your aye vote.

Thank you very much.

Thank you. Anybody else here in support of SB53? Coming up. Name, organization and position.

Thank you, Madam Chair and Members. Megan Subers on behalf of the Secure AI Project, a co sponsor and support.

Morning, Madam Chair. Members Ivan Fernandez, California Labor Fed, in support.

Good morning, Madam Chair. Mario Guerrero on behalf of the University of California in support.

Megan Abell with Tech Equity Action in support.

Todd David with the Abundance Network in support.

Nathan Calvin with Encode AI, one of the co sponsors. In support.

Good morning, Madam Chair. Jai Jaisimha from the Transparency Coalition in support.

Thank you. Anyone here in opposition to this Bill, come on up. Same drill, two minutes each, when you're ready.

Good morning. Ronak Daylami, with CalChamber in an opposed unless amended position to SB 53. We want to thank the author and especially the staff for the outreach on this bill. We understand and appreciate the desire to differentiate this process from the one last year, so we do thank you for that.

As a general matter, of course, we appreciate and share in your goal in ensuring the safe and responsible development of AI. Our concern is in how to best get there. As written, SB 53 presents some challenges that could hinder innovation and may not fully address the true risks of AI.

We are still reviewing the bill and its proposed amendments and discussing the implications with our members. But there are several areas we hope that additional work can be done. First, the bill focuses on developer size rather than model risk.

Small entities can build on models that are just as capable and potentially as risky as those from large developers. Governor Newsom flagged this in his veto last year. Size isn't a reliable risk for, proxy for risk, making the bill both too broad and too narrow.

Second, the bill's definitions of catastrophic risk, critical safety incidents, and dangerous capabilities should be simplified and narrowed. Thirdly, detailed disclosure requirements post significant security risks by making developers internal protocols vulnerable to bad actors. We ask that the requirements be adjusted to mitigate risks and simplify compliance.

We also suggest that the enforcement should focus on material violations and include a right to cure. Developers should not face liability for good faith attempts to meet paperwork requirements, especially when dealing with highly complex and rapidly changing scientific and policy matters. With regard to recent amendments, again, we are continuing to review them.

Some of our members have flat concerns. I can say on the audit provision we are heavily concerned. Cost aside, no one understands this technology better than our own developers. So creating a cottage industry for auditing doesn't enhance safety from our perspective.

Moreover, we also do want to clarify that for the record, the RAISE Act that was actually passed did not include the third party audit requirement. So again, we just want to emphasize that we do look forward working with the author.

We think that there is some place to go on some of these issues and are looking forward to working with not just the author, but also the Committee. So we thank you.

Thank you. She's well practiced. She left you all two minutes.

Okay. Good morning, Chair and Members. Robert Boykin with TechNet here today and respectfully opposed to SB 53 unless it is amended. TechNet appreciates the center's continued engagement on SB 53 and its commitment to responsible AI governance. We support the goal of promoting transparency and safety in the development of advanced AI systems.

We recognize the bill is intended as a working draft. That said, we do have a few key concerns we'd like to highlight. First, the bill applies to all models developed by a company, once that company exceeds a specific computational threshold.

This approach doesn't reflect the actual risk posed by individual models and may unintentionally regulate small lotus tools simply because they were created by a large developer. We believe it would be more effective to focus on capabilities and risk of the model itself, not just the size of the company behind it.

Second, some of the reporting and disclosure requirements, such as those related cybersecurity protocols and internal testing or the reproduction of risk assessments, raise serious concerns about exposing trade secrets or increasing security vulnerabilities. We'd encourage keeping disclosures high level enough to support oversight, but not so detailed that the adversarial actors could exploit them.

Finally, many companies are already investing heavily in robust safety practices like red teaming and post deployment monitoring. We want to ensure that SB 53 strengthens that culture of safety. For example, the adverse event reporting framework may be better suited for national coordination given the cross sector and cross border implications.

We look forward to working with the author to help ensure that this bill is both effective and practical in meeting its important goals. However, for the reason stated, we remain opposed unless amended on SB 53. Thank you for your time today.

I turned my mic off. Thank you. Anyone else here in opposition to this bill? Come on up, name, organization, and position. Thank you.

Robert Singleton with Chamber of Progress. We love CalCompute but remain opposed to the rest of the bill. Thank you.

Aodhan Downey with the Computer and Communications Industry Association. We'd like to align our comments with TechNet and CalChamber. Thank you.

Alex Torres on behalf of the Bay Area Council, in respectful opposition at this time.

Thank you. Senator, I think that is a shorter line than we had last year. So kudos to you. Bringing it back to the Dais. Assembly Member Irwin

Thank you very much, Chair, and.

I didn't see you over there.

I was hiding. I want to thank you, Senator, for your dedication. We all know your, how hardworking you are and trying to secure that AI is safe for all Californians.

But I do need to impress upon you that I hope you take the next two months or so to work with the opposition and make sure that you iron out the differences you have.

This is, now I know you were waiting for the report to come out, but asking the Committee to basically have I don't know what it was one week to digest all this material was a lot and I do commend the Committee and the Chair for, for working so diligently. The analysis is 53 pages.

I think that's a record and an incredible amount of work. But it does point out to, I think there's a lot of work still to work out, a lot of work to look at what the opposition is saying and that makes sure that you deal with the issues.

One of the reasons I didn't vote for your bill last year was where we were, you know, we're concerned in California that we have the right balance between safety and driving out innovation. And I think that the governor's message was the same. So I want to see what the bill looks like when it gets to the Floor.

Hopefully, I can still support it. I'm going to be supporting it today. But obviously I think that you, you do need to look at what the governor's message is to make sure that he can support it also.

Thank you for that. And we're absolutely committed to continued dialogue with the opposition. As the folks in the industry know, we have been incessantly pestering them as more and more, you know, proposals and feedbacks for a while now, and we look forward to seeing that.

And some of the issues that the opposition raised are ones that we, you know, seem like we can absolutely have a conversation about. I know there's a disagreement about audits, and we talked about that last night in Judiciary Committee.

And yeah, we also, we tried, you know, we made clear earlier in the year that we were awaiting the report. We actually even put out an outline of what we were looking at very quickly after the report came out.

So we, I appreciate acknowledging the challenging timing that we're in, but we've been really trying to be as transparent as possible, understanding this is a very, it's a time sensitive situation. Congress is doing nothing, literally nothing.

Congress has yet to pass a privacy, data privacy law, which I think is absolutely embarrassing for the United States of America that we do not have a federal data privacy law in 2025. It's actually despicable. And they tried to ban states from protecting kids from deep fake kiddie porn.

And there were unfortunately, industry actors that were helping push the federal moratorium proposal that was fortunately killed. So I think it's all the more important for California to act and we do look forward to working with a broad spectrum of stakeholders. So thank you.

Thank you, Senator. Madam Vice Chair.

Thank you, Chair. Senator, yes, you, you have become the subject matter expert as well as the Chair here. Let me ask you a couple series of questions. So you've been working on this subject now for a couple of years. What have you learned to, the bills that we saw last couple I mean, we're very broad and kind of all over the place.

Not without good intention. It's just. What have you learned to help us understand what we need to do? When people come and speak for or against, whether they're the platform people are just concerned individuals, including myself, what are we speaking on?

What are we looking for to fix or to take care of?

Sure. And on the, on our bill last year was a, it was definitely a big bill. I think it was, I think it was a well constructed bill and that's why it got bipartisan support and got very broad support in both houses.

And frankly, if you look at the Governor's veto message on it, it was by far the longest veto message I have ever seen. Which, and I think that the, and that reflects the fact that I know that the Governor struggled with this.

I think he really saw the importance of what we were trying to deal with and really struggled with what's the best approach. And I really appreciate that his veto message was very thoughtful and it created the working group. And so, you know, what I've learned is that we all are trying.

We all want technology to succeed in making the world better. I think we all want that and none of us. There are some people, there are some people who want to squash it, but I think that's a small minority view. I think overwhelmingly the public.

And we see this from the polling, because the bill last year was polling statewide, almost 80% higher support among tech workers than among the general public, support among Republicans, Democrats, all over.

And that's not because people want to stop innovation, but they want to make sure that policymakers are actually caring about public safety and what the impacts, both positive and negative, could be. And so what I've learned is that there is broad support for taking steps to protect the public while also promoting innovation.

And so I thought last year that requiring safety testing was reasonable since all the large labs had already committed to, to doing that. Obviously it provoked a big fight. But the bill last year, what was interesting about it, it was the debate about it did not have a ton to do with what was actually in the bill.

With all respect to both supporters and opponents, it became sort of a symbol and a proxy where there were supporters who would talk about it in ways that I'm like, that's not what the bill does. The bill doesn't. It's not as big as some people thought it was.

And some opponents were talking about it in ways that made the bill seem much bigger than it was. But I think it was an incredibly important conversation that SB 1047 sparked.

In terms of what do we mean when we say we want to promote innovation, and what do we mean when we say that we want to make sure the technology doesn't harm the public? And we obviously did nothing for many years around social media and we've seen some of the challenges that's created with kids and disinformation.

And so how do we get ahead of those risks here? And what I've also learned is that there are folks in industry who really get it and want to get to a good place in terms of safety regulation. And they, and they're very genuine.

There are other folks in industry that you could propose the most minimal thing and they will oppose it because they just do not want the government, government to be involved in any way in regulating technology. So it's complicated. But I think this, with a transparency approach, I think we're in a good place, understanding that there's more work to do.

Okay. I don't want to take up too much of everyone's time. Just a couple of questions. I think I said this last year and then I saw it in some comments. This year the EU model was to involve the companies, the stakeholders, if you will.

And it took several years to kind of iron all this out and work all this out. Why is it just to maybe just take further what you just said? Some are willing to work with you and some aren't.

You're working with the large platforms, but it's the smaller companies too, that can gather up $1.0 billion with funding and they're on their way to do whatever they want to do with AI.

Do you think we are really rushing this until we get all the players, stakeholders, to really home in on what you're trying to do for a safety reason? And the catastrophic word is kind of frightening. And how do we even know that what you're proposing will control, transparency yes, but will we really know?

Yeah, I think we can get there. Honestly, I think we know what the issues are. There are some disputes about some of those issues, but I think we have a grasp of what the issues are. And I think it is time sensitive. Right?

This technology is moving more quickly than people anticipated, or some anticipated it, but others didn't. And there's all sorts of predictions about when are we going to get to artificial general intelligence, when is there a possibility of some sort of super intelligence. And people have all sorts of different predictions on it.

But just in terms of generative AI, it is moving very, very quickly. And those impacts, we're seeing it with job impacts, which in some ways is the hardest issue, the jobs issue, because technology throughout human history has had impacts on how people work and what jobs exist or don't exist. It's never happened at this speed.

But then on safety, on AI making decisions about your life, which could be profound decisions about your life, about employment, about banking, about so many other issues where all of a sudden a decision is made about your life that impacts it in a grave way and no human being was involved.

And so there are just all sorts of issues that we're grappling with. I think we know what those issues are. There are disputes about how best to address it, but it is time sensitive and I think having laws with some flexibility in it is a good thing. And we try to do that in this bill as well.

I'm just concerned that they're not working with you.

Oh.

I mean, I think.

I think they will.

It's been going on for some.

Some folks in industry have been working with us. We've been in conversations with the Chamber with, with a large laugh. We've been reaching out to the.

What about the small? The small.

Yep, we, we've reached. Yep. We've been, we've been talking to everyone. Yes.

Well, it would be good, I mean, if there could be a coming together, a common voice for where we are right now. Now, of course, the Chamber, the business community, as you said, you're not, you're. You would support amendments and fixes. It's, it's kind of, it's still in work, work in progress.

I, I accept that. Is? Okay. So does industry see this as the definitive policy that is going to move forward with some.

I can't speak for industry, but I can say is that California has an outsized voice. And when the Federal Government is not apparently going to do anything to regulate it in any way, which seems likely, at least in the near future, California has an incredibly powerful voice because of just.

We're the capital of tech innovation globally, and we still are, no matter what Austin or New York might say where, it's still here. We're proud of that. And so when California speaks on an issue like this, it has international reverberations.

I can tell you the number, the amount of international outreach I got about the bill last year, it was extraordinary, constant. They were all watching what we were doing in California.

And one final question, if this is even answerable all these bills that are pending now before the Committee and moving through the process related to privacy, related to children and well, all the. Would this bill resolve that so we don't deal with chat boxes creating horrible things? Would this bill take care of that?

So we'll still have those kinds of bills specifically addressing various problems?

Yeah, this bill is about safety for really significant safety risks. Other bills may be about discrimination or about deep fakes.

So that will continue.

That'll continue.

This is kind of the foundational.

For one piece of it, the significant safety risks.

Okay, thank you.

Thank you.

Sorry to prompt you in the middle of that, but I think it's important because I've had the privilege of talking to some of the industry players that the Senator has been consulting with on this.

And I think, but correct me if I'm wrong, opposition up to this point, you were not opposed to the prior to the Senate version. Am I right about that?

We did not have a position on the prior version. We were talking with the author's office quite extensively since even before the introduction they've done outreach to us.

So if it's helpful, if I may, we, we represent both small and large businesses and certainly we need to be at the table and have these conversations because it affects all of our members and it, it wouldn't behoove us to not be having these conversations.

So we're happy to have them and hopefully we can get somewhere good this year.

Yeah. Okay. Assembly Member

Thank you, Madam Chair. And thank you, Senator. I mean, one thing that's really reassuring for me is we're all on the same page about something, that something needs to be done. I supported this bill in Judiciary because of the sheer volume of information.

I do want to echo the concerns of my colleague and also tell her thank you. Because I do not come from the tech world and because of the Tech Caucus, I was able to visit what I believe is your district and toward the extraordinary world that is the tech industry in the Bay Area.

We need to make sure that we're supporting them and keeping them there. And I think that's exactly what you're trying to do and making sure that we're protecting the public at the same time. So I'd love to be included in conversations with this because I, it's just this world is so extraordinary.

We are in like the horse and buggy era of AI and we could be in flying cars in the very near future. But I want to make sure that the right safeguards are there.

So I'm not going to be voting on this today, But I do look forward to supporting it on the Floor because I know how hard working you are and willing to get this exactly where it needs to be. So thank you very much.

Thank you.

Thank you. I don't see anybody else. Is there anybody else? Okay. Well, I want to thank the Senator for working with our staff and also to the staff.

I think I appreciate some of Irwin's comments because they are a small but mighty team with the number of bills that come through this Committee these days and the complexity that we see. And I think the Senator has been collaborative both with Committee and also with many stakeholders as this moved forward.

And I think there are a couple things that are significantly different about this year that I want to highlight in addition to what the Senator said, which is under this bill, one of the things I hear from industry a lot is we, and I think I heard you say this, we know our products best.

We know what we should be doing. We know how to keep, how to create safety protocols. And that is fundamentally the basis of this bill is that they will be defining their own safety protocols. They will be making those public, and then they will have to follow them.

And that feels like that is significantly different from last year. And it also is putting the responsibility on the AI companies to decide how to keep the community safe, but allowing us, as the people whose sole role is to keep communities safe, informed in how that progress is happening.

I also appreciated one of the amendments that the Senator's taking here, which is that for. I mean, the risks that are contemplated in this bill are risks to human life. I mean, this is the most fundamentally important thing that we do, I think, is keep communities safe.

And so one of the amendments is that within 24 hours, the company would be obligated to tell the relevant governmental authority that could help keep communities safe in the event that something were to actually come to pass.

And I think that's an important amendment because what it shows is that as this innovation happens, we are just asking that you bring us the government, police, our public health officials, it depends on what risk is coming to pass, into the conversation to help keep communities safe. And I think that's so fundamental to our role.

And you are nothing if not tenacious. So I know that you have stuck.

And tall.

And tall, you have stuck with it. I think, because you believe so deeply, despite the many criticisms you got for this, that this is really important.

I'll also note that since last year, New York has already passed a law that is different, but is similar in a lot of ways. And so we are not the first anymore because of the veto.

And so I do think this is something that two of the largest economies in the country are starting to look at is how do we keep our communities safe as this innovation happens. I've always been a fan of CalCompute, as you know.

I have told you that I'm happy to be a partner in trying to see that come to fruition, because I think competition in the marketplace is also one of the things that will keep us safest as AI comes to pass. And that allowing, democratizing Compute is one of the ways to do that.

And currently you have to get $1.0 billion in order to get into the game. And the idea that we, as the world's capital of innovation, would say no, we're going to let everybody play this game and we're going to let everyone try to innovate and create miraculous things here in California.

We're going to provide the resources to do that, I think is something that will keep communities safer. So built into this is really important pieces that I think come together to create a safer ecosystem.

Also, the whistleblower protections, which I know you weren't opposed to because those were in the bill prior, are really important because currently you can only whistleblow for things that are illegal. And a lot of what this bill contemplates isn't.

Actually nothing has reached the point of illegal conduct when we need to know there are risks coming to pass. And so that's an important change to whistleblower law that I think is critical and will keep our communities safer. And so I'm happy to support it today.

I'm happy to hear that we have bipartisan supporting concept, if not a practice. And with that, would you like to close?

Thank you so much. I appreciate the dialogue today. You know, this is a bill who I think it's, I think it is important for us to take action this year. We know that there's going to be additional dialogue. It's a different approach than last year, and I think it's one worthy of support.

Appreciate the Chair and the Committee working with us. I look forward to the continued work, and I respectfully ask for an aye vote.

Thank you. And we have a quorum, so let's call the roll or not. Let's. Yes. You know what I mean.

[ROLL CALL]

A motion and a second. Let's call the roll.

[ROLL CALL]

That Bill has seven votes. We'll keep the roll open for absent Members. Thank you, Senator. We have. Senator Allen. Beautiful. SB766. Sorry. oh, yes. We have a motion on the consent calendar. Do you have a second? Wait a second. Call the roll in the consent calendar.

[Roll Call]

He left.

[Roll Call]

Recall when you're done.

[Roll Call]

Recall. Lowenthal. He's here. Lowenthal, he's right there.

On the consent calendar, Lowenthal.

Mr. Lowenthal. On the consent calendar.

[Roll Call]

Perfect. Sent. We will. We have seven votes on the consent calendar. Will keep the roll open. Senator Allen.

Madam Chair, congratulations on a marathon close last night at the Judiciary Committee with lots of tough but important privacy related bills. So consumer. Yeah, sorry. That all makes sense now. Yeah.

12 hours later, it all comes together.

I mean, only barely, but. All right, well, let me accept the Committee amendments to SB766 and I'm, you know, happy to announce these amendments are expected to move our opposition to neutral, which is very exciting.

So this Bill seeks to add consumer protections, important consumer protections for car purchasers and car leasers by codifying the FTC's cars rule and establishing a cooling off period for used cars. Late last year, the Court of Appeals for the 5th Circuit overturned the cars rule, but only on administrative grounds.

We had many Californians filing comments in favor of the FTC's rule that this Bill codifies, including former auto sales personnel, consumers who personally experienced auto scams. Our Attorney General and basically this was a look at the pricing and advertising of cars and making sure that we try to refine our rules to make them more consumer friendly.

So what does the Bill seek to do? It would prohibit car dealers from misrepresenting the price of a car, require transparency on the full total price and total payments. It requires the dealers to be transparent about add on services or features that are optional, prohibits them from charging for add ons that will not benefit the consumer.

And then it will provide additional protections for Members of the military and veterans by prohibiting dealers from misrepresenting affiliation with the US Government, state government, local government, or any agency or Department, including the Department of Defense. Additionally, there are some limited opportunities for consumers to return a car under current law.

Existing law uses allows for car purchasers, used car purchasers, to have the ability to return their car only if they pre purchase the right to return it.

But this requirement is a strong deterrent and many purchasers are unable to make the additional purchase when they're buying the car so as to give consumers time to review the purchase agreement carefully and fully inspect the vehicle.

The Bill as proposed via amendment, creates a three day cooling off period that doesn't require consumers to pre purchase the return period, and it would permit dealers to charge a restocking fee. The cooling off period includes some additional guardrails.

By capping the maximum value of the used car at $50,000, it also caps the mileage a consumer may drive so that folks don't go on some joyride to Vegas and just effectively rent it for free for a couple days.

So here with me today to testify in support of the Bill is Rosemary, who's on behalf of Consumer Survival, Reliability and Safety.

Madam Chair, two minutes each when you get a chance.

Thank you, Madam Chair. Members, Rosemary Shahan, President of Consumers for Auto Reliability and Safety. And while we're not the sponsors of the Bill because the author is the sponsor, we've been very active and enthusiastic supporters. We applaud Senator Allen for authoring this Bill.

According to the Federal Trade Commission, just the parts of the Bill that emulate the FTC cars rule are projected to save car buyers $3.4 billion a year and another 72 million hours that they would otherwise spend haggling over the price of cars. So that's a real benefit right there.

And then in addition to that, the Bill has provisions, as the Senator mentioned, for a kind of mini auto escrow period during which at least some consumers will be able to get the car inspected and it will benefit consumers like William Bradley, who testified before the Bill in Senate Judiciary, who bought a car, paid 10,000 plus in cash, was told that although the Carfax report said it had come here from Pennsylvania, that rust was not an issue.

And then shortly afterward, he got the car inspected and found out it was so badly rusted someone had taken duct tape and taped the undercarriage together that it had major structural damage and made it grossly unsafe. The auto tech said, I'm not letting you leave the lot in this car.

But when he went back to try to get his money back from the dealer, the dealer said, well, you bought it as is, so you're stuck.

And I will be thrilled if the Bill passes, to be able to tell William Bradley that the next time he goes to buy a car, a used car, for under $50,000, if he has a similar experience. He can get a refund without having to get an attorney. It should be a relatively painless process. That's groundbreaking.

No other state has done this before. So we really appreciate the chair and the council and their role in bringing about agreements.

I'll be honest, we're not going to tout it as like the model because we want to like take it for a test drive, maybe give it a tune up before we urge other states to adopt similar legislation.

But we really appreciate all the hard work and especially I want to recognize Shoshana Levy, who's been amazing to work with and incredibly just smart, great and hardworking. So and with that I'd ask for your aye vote.

Thank you.

Good morning Madam Chair, Members. Robert Hurrell. I'm the Executive Director of the Consumer Federation of California in strong support. We've been intimately involved all year along with a few other groups in the negotiations. This is a purchase where there's a major power imbalance between the consumer and the entity they're purchasing it from.

And so we wanted to make sure that to try to. You'll never be able to fully level that playing field, to be honest.

But what you can do is empower consumers to have more opportunities and rights to make sure they're not misled and make sure that if they didn't get what they thought they were getting, that they can go back. So I just want to thank chair and your staff who worked with us intensely over the past week or so.

Thank the author. His staff has been fantastic and the whole team that's been negotiating on this and we urge an aye vote. Thank you.

Thank you. Anyone else here in support of this Bill? Come on up. Name, organization and position.

Good morning. Danny Kando-Kaiser with Kaiser Advocacy here on behalf of the National Consumer Law Center, California Low Income Consumer Coalition and Consumer Reports in strong support.

Good morning chair and Members, Sosan Madanat, W Strategies here on behalf of Unidos US in support.

Thank you.

Good morning. Cristel Chavez with the Mesa Verde Group on behalf of the Consumer Federation of California in strong support as a co sponsor. Thank you.

Thank you. Anyone here in opposition or tweeners? Since I know we have a bit of a no longer opposition situation.

Not, not, not a tweener. Kenton Stanhope, Chair Members of the California New Car Dealers happy to report with the July 14th amendments and then the amendments outlined in the analysis, we will be going from opposed to neutral.

Do want to thank the author and his staff, Shoshana, for all their hard work and a really big shout out to the Chairwoman and especially Committee staff Josh Tosney, we wouldn't have landed this plane without your leadership and hard work. It's not even a question. This is the most massive overhaul and consumer protections of all 50 states.

And it will be a model for the rest of the nation. I have no question about it. Thank you.

Thank you.

Good morning, Madam Chair Members. Mandy Isaacs-Lee here on behalf of Carvana. Also echoing our thanks to you, the. Chair, the Committee, the. The author and the staff, we are pleased to remove our opposition as well.

Thank you.

John Moffatt here on behalf of Enterprise Mobility, similarly situated. Thank you.

Thank you.

Neil Clark on behalf of Insurance Auto Auctions, also moving to neutral. Thanks for working with us.

Yeah.

Madam Chair, Carlos Gutierrez here on behalf of CarMax, also removing our opposition, moving to neutral. And thank the Committee and Senator for all his work. Thank you.

Ashley Hoffman, the Cal Chamber, also moving to neutral, thank you.

Madam Chair. Chris McKayley on behalf of the Civil Justice Association of California, moving to no position. Thank you.

Thank you. Seeing no one else in opposition or any other status. We'll bring it back to the dais. Yeah, we have a motion. Madam Vice Chair.

Well, I'm sitting here in pleasant astonishment. So you, you have brought the opposing. Forces together, if I'm understanding these comments together. So I guess I am surprised that. In the State of California there has. Not been a used car consumer set of laws or law ever has there been. Is this the first?

No, no, no, no.

But we are need to use car sales and returns and that type of thing. But this is, this is new nationally. Is that what I just heard?

We're, you know, certainly we're proud of the, of the, of the progress we're making with this Bill in terms of consumer protections. And I think, you know, we'll see what impact this has nationally. Obviously, you know, the bill's not as strong as it was when we first introduced it, and yet I still believe it's very strong.

And so, I mean, depending on the, I mean, I'll turn this over to our friends who work in the consumer space to answer that question.

Okay, that's a really good question Assemblymember. And there is a California Car buyers Bill of Rights that was enacted in 2005 that was a legislative compromise based on ballot initiative where I kind of remember.

And we ended up with, instead of a cooling off period, we ended up with a two day return option that consumers had to purchase in advance and, and nobody was getting it right. So this is very different because this is going to be a three day right. Consumers will have that right.

And in other states, very often there are used car lemon laws that mandate that dealers have to offer warranties of various lengths, depending on the cost of the car or the mileage.

We thought instead of that, which you often have to enforce by getting an attorney and going to court, a cooling off period where you could just say, you know what, this isn't working for me.

I'd like to take it back and get a refund would give people the opportunity to get out of a bad deal for whatever reason without having to go through the legal system, not even going through arbitration, but trying to make that as easy and painless as possible.

Well, it makes a lot of sense. In that you have brought the stakeholders. The new car, the used car dealers, the consumer group together, if I'm understanding that correctly. Yeah, well, well done. All right, I'll be supporting the Bill and I'm really pleased to see that. All forces came together to do good.

Work for our consumers, for all of us. We all will benefit from this. Thank you. Thank you.

Thank you so much. Thank you. Anybody else? Okay. Well, I want to thank you, Senator Allen, and also to your staff and to the staff of the Committee. Often we hear even in what is the Last Committee? Oh, we'll work on it later.

And I always think by the Last Committee, we should have bills in a state that they are ready to be sent to the floor. And you have done that work, and that is clear.

And I think it is commendable because you set out with the goal of protecting consumers from, to your point, what can be a really complicated and large purchase for many people, this is, you know, they're not homeowners. This might be the largest purchase they make.

And so making sure that they get what they are buying and intending to buy is critical. And so I agree with the now neutral former opposition that I actually do think this is a model. I think you are pushing the boundaries for consumers further than we've seen anywhere in the nation.

In the fourth largest economy in the world. You are protecting 40 million consumers. Well, not so many because some of those are kids, but a lot of consumers. And that's really commendable.

And I think you should be proud of the product that's coming out of here that will have bipartisan support, that will, I believe, be hopefully signed by the Governor become law for Californians. And so I commend you on that work, and I think it's work I've seen you do over the years.

But you continue to hold strong to your goals and also work with people to make sure it's workable. And that is commendable. So with that, I'm requesting an aye vote. And would you like to close?

Yeah. Thank you, Madam Chair. Look, you're right. I mean, this is a hard process. You know, this was a much. I'll tell you, the first hearing was not anywhere near as pleasant because we wanted to make real change in this area.

And yet, you know, our friends in the opposition have, you know, important business model that they seek to protect for all sorts of good reasons.

And, and so there's so many thank yous to give, and I hope you just indulge me, first of all, Madam Chair and the Committee staff for helping it to really bang heads together at the end. And I think we were getting closer and closer through Committee, through Committee. But to land this deal was incredibly important.

And your leadership in that effort was really, really valuable. And I just continue to, you know, your staff is legendary. You know, it was at the heart of a bunch of the tough bills last night, too. I know.

And just want to recognize that Robert and the folks at the Consumer Fed, Rosemary, who's such a legend in this area, Rosemary Shahan, who's been working so hard on consumer protections in this space for so long, just thank you for sticking with us through this whole process.

I know it wasn't easy, and you've been our kind of moral guiding star in this effort. Obviously, my friends in the opposition who have now gone to neutral, you know, thank you for working so hard with us to get there. So many of you I see in the room and couples that aren't here, Anthony and others.

But just appreciate, appreciate the hard work we all did together to get this landed. And then of course, I do want to just especially highlight the extraordinary work of Shoshana Levy. So much of our work Members is done by our staffs who make us look good. And Shoshana, Rosemary is absolutely right. Brought intelligence, determination, patience.

I know she spent just this last weekend entirely working on this. And it was just the most recent of many, many weekends. So I just want to thank you, Shoshana, for your hard work on this really important Bill that's I think, going to make a big difference for consumers. So thank you, Members.

Thank you for recognizing all the work that went into this and everyone coming together for something I think is going to be good for fellow Californians. So with that, I respectfully ask for an aye vote

Thank you. And if you'll indulge me, also, I want to thank the. I didn't. And I failed to mention that the former opposition, I mean, they really came to the table.

And I do think that is what we should hope for in this building, is they, you know, they showed that they cared equally about their consumers and that they wanted to move the ball forward in a way no one else had. And I know it wasn't always easy, and it wasn't.

Didn't land exactly where it started or where maybe you would have wanted, but they absolutely came in good faith and worked really hard. And that is important to commend that, and I hope more opposition will do that.

And we still really care about where the Bill is. Yeah. So.

Yes. So we have a motion. Do we have a second that.

[Roll Call]

Thank you, Members. Appreciate it.

Thank you. That Bill has seven votes. It will stay on call until we get our Members back that are all over the building. We have no authors, Senators, so come on down. Senator Wiener and Senator Allen were so prompt this morning, and we're here at 11:30 at night, so.

They really are. Senator McNerney's here. Come on up. Senator McNerney, this is SB7.

I wasn't sure if you were in recess or not.

No, we were just waiting for you.

Okay. Proceed when ready. Well, good morning, Chair Bauer Kahan and Vice Chair Dixon and distinguished Members of the Committee. I'm pleased to be here this morning. It's still morning, isn't it? To present SB7, I want to start by thanking the chair and the Committee staff for working with us on the amendments.

We've gotten a lot of input over the last couple of months, and we've made our best effort to incorporate that in a way that it doesn't kill the intent of the Bill. SB7 puts in place common sense guardrails. Little out of breath after running over here. Excuse me. What's that? No, no, I'm good.

But the common sense requires common sense guardrails. It requires Employees to notify workers before using automated decision making systems in employment decisions. Number two, it ensures that there's a human in the loop for discipline and termination decisions. And three, it bars ads, automated decision making systems from predicting workers future behavior to justify employment actions.

We have worked again extensively with all the stakeholders as a result of those conversations. The Bill is narrowly focused and much more workable while keeping strong protections in place. So that's my opening salvo. I'm going to turn it over to my witness, Ivan Fernandez from the California Federation of Labor Unions and then we'll start the discussion.

Thank you.

Good morning, Madam Chair and Members of the Committee. Ivan Fernandez, California Federation of Labor Unions proud co sponsor of SB7 automated decision making systems or ADS as this Committee is very aware, may be a new tool employers have at their disposal and to squeeze out every bit of productivity from a worker. However, this practice is nothing new.

Employers have used scientific management to speed up, monitor and control workers for decades. Today, employers are outsourcing management to AI powered systems capable of firing and disciplining a worker and determining how much they are paid. For example, Amazon Warehouses use an unpaid time system that provides workers with a limited amount of non working minutes throughout the day.

However, rather than providing that said flexibility, workers have reported that the system is used to automatically generate paperwork to fire them if they use up a certain amount of minutes of their allotted time. Workers can be fired even if they're 10 minutes late.

Because of circumstances like a sick child or a traffic jam, ADS is incapable of making those nuanced decisions. Another risk of unsupervised ADS usage is the potential for bias and discrimination built into these systems. Technology is not free of biases and the flaws of the humans that create them.

Human biases can also be built into the core of the technology through training, data and the algorithm itself. This risk is worsened when these tools are used to predict future behavior rather than just analyzing past performances.

Just being flagged as a risk can impact workers ability to get good shifts arrays or put them at risk of termination simply because of an algorithm. SB7 requires human oversight of machine management with the requirements as described by the Senator and and for these reasons we strongly urge your aye vote at the appropriate time. Thank you.

Thank you. Anybody else here in support? Come on up. Name, organization and position.

Yadi Younse with Oakland Privacy in support.

Wait, the mic's not on. Come back. Is the mic on? Try again.

Yes. There we go.

Can you hear me?

Yadi Younse with Oakland Privacy in support.

Thank you.

Megan Abel with Tech Equity Action in support.

Good morning. Connor Gusman, on behalf of California Teamsters, proud to co sponsor this Bill as. Well as Unite Here, United Auto Worker. Or sorry, the Utility Workers Association, the International Association of Machinists and the Engineers and Scientists of California. Thank you very much.

Thank you

Madam Chair. Members. Robert Hurrell with the Consumer Federation of California in strong support. Thank you.

Thanks.

Good morning, Madam Chair. Jai Jaisimha from the Transparency Coalition. As someone who worked on this type of tech at Amazon, I'm in support of this Bill.

Thank you.

Good morning, Madam Chair. Members of the Committee, Navnit Puryear on behalf of the California School Employees Association in support.

Thank you.

Good morning. Ariana Montez, on behalf of the Communication Workers of America, District 9 in support.

Thank you. Okay, you're coming up for opposition. I thought maybe you were coming up one more support, but you guys can come up

AFSCME strong support. Thank you.

Okay. Thank you. Oh, I appreciate that. We appreciate speed. If anybody remembers how quickly I spoke when I was presiding officer, we appreciated it.

Good morning, Madam Chair Members. Chris McKayley, on behalf of the Society for Human Resource Management. As the Senator knows from our testimony in his house as well as earlier here, we do appreciate the continued engagement of him in the Federation of Labor Unions.

We want to be clear, we're not opposed to providing employee notices with ADS that are used in significant ways in the workplace.

Our concern is first that SB7 requires notices to both employees and independent contractors when ADS is used at all, no matter how minimal it is, for example, including scheduling, task allocations, et cetera, with particular concern about independent contractors.

Ironically, in the definition of worker in the Bill, it talks about direction and control, which of course is a traditional definition of employee, but later in the definition of worker, it includes independent contractors. The second is in Section 1522, subdivisions A and B.

The length of this notice that's outlined in those two provisions, we think, particularly for a large employer, is going to be very significant and fear that it will turn into what many of us view as Prop 65 warnings. We simply bypass them.

The other item is in proposed Section 1524, where it includes a complete ban on predictive behavior analysis. And we think that that incorrectly assumes that this is always used for some sort of nefarious purpose.

In many instances, for example, a financial institution that is trying to combat consumer fraud, obviously their employees are going to have access to consumer accounts. And so even if a flag is issued by an ADS and is fully investigated, that initial flag could be unlawful under this Bill.

So for these reasons, we continue to maintain an opposed unless amended position on the Bill. Madam Chair, thank you.

Thank you.

Good morning. Ashley Hoffman, on behalf of the California Chamber of Commerce, I want to echo the thank you to the author and staff and the sponsors for all the conversations on this Bill.

And what we've said from the beginning is we absolutely agree that a human should be in the loop when you're making consequential decisions like a discipline or a termination from a labor and employment perspective, that absolutely makes sense.

I think one of our concerns here is that the Bill, especially the appeal right, for example, is focused a lot more on what happens afterwards, not ensuring that we're having that human in the loop element up front. Again, I think that's really where we think the focus of the Bill should be.

I know there's a recent amendment to the appeal right and I've asked but haven't gotten complete clarification on this. But to me it still reads that even if an ADS is used at all, no matter how minimally, that would trigger the appeal. Right. Even if there was, for example, a thorough investigation performed.

And so we do have some concern from that perspective of the precedent and again implications that could have on broader labor and employment law. We are also concerned still about the provision stating that ads can never be relied on, quote, primarily for disciplinary purposes, no matter how impeachable it is.

How much testing, for example, you've done on that. And one of the issues we have implementation wise is workers always be observed at all times. And so we have to rely on tools, for example, sometimes for safety in the workplace if the same worker is repeatedly flagged by a tool for unsafe driving.

As I read this Bill, I am not sure that my Members could discipline for that. We would, for example, have to have someone maybe drive along with them constantly or some other means or measurements of productivity.

And I know that can get controversial, but there are some legitimate uses where ads can flag if nothing is being done by a worker or again, there's potential risk of unsafe behavior. So again, we're not opposed to level of corroboration, have some concern about the use of the primarily language in the Bill.

And then finally is the access to access rights and the correction rights. As I read this Bill, it applies to all low risk uses because it is governed by that definition of an employment related decision which covers things like scheduling, task allocation.

As my colleague mentioned, I know the proponents in the analysis said our argument on this doesn't stand because they assume a worker is not likely to request or access data unless there's appeal, but again, it's not actually that limited.

So again, you know, we really believe that the better approach here is again, more focused on kind of having that human involvement up front. I know. We've sent over red lines and continue to work with author. Thank you.

Thank you. Anyone else here in opposition, Name, organization and position, please come on up.

Good morning. Jacob Brent with the California Retailers Association in respectful opposition.

Good morning. Robert Boykin with TechNet, in respectful opposition.

Morning, Madam Chair. Dylan Hoffman, on behalf of Public Risk Innovation Solutions Management or PRISM, respectfully, respectfully opposed. Pardon me.

Thank you, Madam Chair. Members Jean Hurst here today on behalf of the Urban Counties of California, the Rural County Representatives of California and the California State Association of Counties, opposed for the impacts on the Public Safety Center.

Robert Singleton, Chamber of Progress, also respectfully opposed.

Margaret Gladstein, on behalf of the Security Industry Association, respectfully opposed.

Good morning, Madam Chair. Aaron Avery, California Special Districts Association, respectfully opposed. Also, on behalf of my colleagues at the Association of California Health Care Districts. Thank you.

Thank you.

Sarah Bridge, on behalf of the California Manufacturers and Technology Association, we align our comments with the Chamber of Commerce and we respectfully oppose.

Thank you.

Good morning. Kalyn Dean with the California Hospital Association in opposition.

Good morning. Chair and Members. Robert Wilson, California Credit Union League, here in opposition. Thank you.

Good morning. Matt Easley, on behalf of the Associated General Contractors, both the California and San Diego chapters, in opposition. Thank you.

Debra Carlton with the California Apartment Association, in opposition.

Thank you. I'm seeing no one else here in opposition. We will bring it back to the dais. Questions? Oh, yeah, Assemblymember Irwin.

Well, thank you for bringing this forward. I think the most important thing is. That. We're prudent and put a person in the loop when these automated decisions are made. But could you answer a couple of.

I mean, we definitely heard a lot of issues, a lot of opposition brought up, but could you talk specifically about banning all predictive behavior? There could be, again, some instances where being able to predict behavior could be useful to protect employees. And I think that that was mentioned with the looking at the speed of a car.

And then also, if you have a person in the loop, is it redundant to now say the employee can still appeal the employment decision?

Because I see it and I know this was all brought up by the opposition, but it might be for a small business that the person that was in the loop might be the only person there, and you have to get a second person that the employee could appeal to.

So would you mind just addressing some of the items that were brought up.

Well, I think I'll address the second item first is having a human in the loop. It's kind of it's going to be difficult to get that exactly right, but we've worked with both sides of the of the issue and we've come down pretty well on that.

So even if there is a human involved in overseeing a process, then there's mistakes that can be made, especially if the person has too much of the automated system.

So in other words, if you have a person supposedly overseeing, tasked with overseeing and an automated decision making system making a decision without really that much oversight, then I think mistakes can be made and people can be harmed. So I think it's important to have that necessary human in the loop part of the process.

And the other question, I think I'm going to turn that over to my witness Ivan

Thank you for the question. Assemblymember. So regarding predictive analysis, you know, we understand the need and the desire for employers to analyze data and we definitely don't prevent that through this Bill. Where predictive analysis can get difficult is there can be a variety of pieces of information.

So for example, if an employer wants to determine whether or not a worker is likely to commit fraud, they can put, let's say, their credit score or let's say their income level or where they live and all of these things can or these metrics or pieces of information can further exacerbate biases of who is likely to commit what type of, let's say, crimes or actions in the workplace.

So our hesitancy is to make sure that those biases aren't exacerbated by predictive behavior analysis.

No, I totally get that part. And in support of that, I'm just talking about if you have a predictive analysis that would predict that an employee is likely to get hurt because they're driving too fast, for instance. So are you preventing, are you preventing that type of analysis by banning that predictive.

If I may through the chair respond, no. So what in those instances, like in the driving example, the employer would have all those logs of, you know, the system catching that person going over the speed limit, breaking the law while they're driving. And so the employer themselves can determine, well, they've broken the law a few times.

In these instances, that itself warrants, you know, an action from the employer to the worker. So it's not necessarily eliminating that ability to say you've already broken the law a few times. Let's take an action. Let's, you know, you just can't use.

That to predict future behavior is what you're saying?

Thank you, Assemblymember.

I just I also want to mention that we're still open to finding language on the issues that you've raised to improve the bill. It's always important you want to put out something that's going to be effective and not have undesired consequences. So we're certainly welcome. Additional comments from the opposition from the Committee.

Thank you, Senator. Anybody else? Okay. Seeing no further questions or comments. I want to thank the Senator for his hard work in this space. You know, it's a space you've decided to really dig into.

And I think that your expertise and experience is lending itself to important work protecting Californians and here specifically, workers in a new age of technology. And so I know how hard you've worked meeting with the opposition repeatedly, as I understand it, and continuing to listen. And I know that work will continue as we move forward.

So with that, would you like to close?

Well, I would first of all want to say how much I appreciate working with the opposition and with the staff and Mike in my office and the staff here in the Committee. It's been very helpful. It's still an open dialogue, but I think we're pretty close to where we want to be on this.

And I will respectfully ask for an aye vote.

Thank you. And do we have a motion on this? Not yet. We have a motion and a second. Let's call the roll.

[Roll Call]

That bill has four ayes, two noes. We'll hold it open for absent Members. More people will arrive. And now we will take up Senator McNerney's second bill, SB 833, when you're ready.

Well, another good morning. And again, I want to appreciate the Committee for hearing this bill. I want to start by thanking the Committee and accepting its amendments and the Committee staff for their hard work.

The amendments in this bill remove the adverse event reporting provisions to avoid conflict with another author's bill and to include another provision to include flexibility into the bill to ensure that the human oversight doesn't disrupt agency operations.

This bill, SB 833, ensures human oversight of an AI in an artificial, in a critical infrastructure by requiring real time monitoring and human review before AI can act. It also requires the Department of Technology to provide actual annual training and conduct system assessments. This bill keeps a human in the loop before AI makes critical decisions.

It's aligned with the Governor's executive order and international best practices. It's a simple, straightforward way to prevent avoidable disasters while allowing innovation to continue. With me today in support, I have Jai Jaisimha from the Transparency Coalition. ai. Jai.

Thank you, Chair and Members of the Committee. My name is Jai Jaisimha. I'm testifying in support of SB 833. SB 833 establishes human oversight requirements, which I think are critical. If you're wondering whether it's time to act, I gave you this from Monday's MIT Technology Review.

California is set to become the first US State to manage power outages with AI. I'm sure it comes as a surprise to many people here. It certainly came as a surprise to me.

And the use of generative AI in this environment is, without proper guardrails, is going to be potentially very problematic. Without proper oversight, assessment and transparency mechanisms, AI systems operating in these critical sectors could make consequential decisions without human review or worse. There's also concern that unmonitored AI could take risks that could potentially escalate.

In this particular example, the one with the article I just cited, they're planning to use generative AI. So you can imagine all the issues that you experience with generative AI around hallucinations, all of that potentially being in place without any safeguards or oversight.

SB 833 would go a long way to address these issues and improve the safety of AI systems. I urge the Committee to stand up for California and pass SB 833 with an aye vote.

Thank you so much. Anyone else here in support of this bill, come on up. Name, organization, and position.

Yadi Younse with Oakland Privacy, in support.

Thank you. Anyone here in opposition to this Bill? Seeing none. Oh, wait. I was so excited.

Good morning, Chair and Members Robert. Boykin With TechNet moving from an opposed to an opposed unless amended on the bill due to the amendments being taken today. We have other concerns regarding narrowing the bill to high risk ADS. We appreciate the author's willingness to work with us and we look forward to continuing conversations if the bill moves forward today. Thank you.

Thank you, Chair and Committee. Ronak Daylami with Cal Chamber. We are also reviewing the amendments. We'll likely be changing our position, but we will follow up as soon as we get a chance to review them in full. Thank you.

Aodhan Downey with the Computer and Communications Association. I'll just ditto what they were saying. Thank you.

Thank you. I'm a little confused by the opposition. Maybe I'll make them come back up. Anybody have any questions? Questions? I have a question for the opposition. Did Robert want to come back up? Did he leave?

Don't see him.

I don't see him either. I guess I'm a little perplexed if somebody from the opposition wants to. Yeah, Robert, come back up. We have a question. I have a question. So help me understand.

I mean, what the Senator is trying to achieve here is making sure that for critical infrastructure we retain a human in the loop that creates policies to ensure that these function the way we want. This is our most critical public infrastructure.

The amendments he's taken in this Committee personally, I, having worked so deeply on 911 and 988, understood that we needed to give more flexibility to agencies. You can't have a human deciding every time 911 uses an ADS that would make 911 fall apart. Right?

So we have to defer to agencies to allow them to decide how to keep the human in the loop. But I think fundamentally this is a really simple bill. I actually was surprised there was opposition. Can you give us a little bit more of why you would be opposed to this?

Oh, it's just some concerns. Wanted to narrow the scope. We think it should be applied to high risk situations. More, more high risk situations than the day to day ones. It's just an ongoing conversation we're having.

Isn't critical infrastructure by its nature always high risk?

Maybe we can say that. I'll go back and forth with that, but that's the, That's. I don't know that for sure. And I will go back and take that, but I don't. I was asked. Sorry, caught me off guard.

Well, if you're gonna show up and oppose the bill, be ready. So. Okay, thank you. Okay. Well, I just am confused because I feel like what you're trying to do here is, make sense to me.

I was obviously, I actually think that the most important people to be talking to in this is not confusingly the opposition that showed up today, but the agencies. Like will this work for them?

Which I think was really the conversations that we had in preparation for this because I think, you know, to me whether our grid stays up is obviously critical, for example. So, you know, I think you were focused on this and making sure it works for the agencies.

But I imagine, and I hope that the government, that the government agencies will be on board with the idea that as we deploy these ADSs there needs to be someone that is a human that's overseeing them. And so I absolutely support this effort. Senator Wilson had some comments or questions?

I've been running back and forth between committees. I wanted to be clear, is this SB 833 or SB 7?

833.

Okay, then I missed.

You missed SB 7.

7. All right, never mind. I'm good.

And she's gone. Okay with that. I don't see any more questions or comments. Thank you. Sorry, my mic was off. Thank you, Mr. Boykin, for coming back and indulging the Chair. Do we have a motion on the bill? We have a motion and a second. Perfect. Let's call the roll.

So I don't need to close, I guess.

Oh, sorry, I forgot the close. Senator, would you like to close?

Again, I thank the Chair. This is a common sense bill. We've narrowed it down really quite a bit. We don't want to cause the agencies to have problems and we've given them quite a bit of flexibility.

And you know, there was a report by, on a test by Anthropic that showed that AI systems would be willing to lie, cheat, steal, and do anything they can to protect themselves, including harming humans. So it's needed. This is a fairly narrow and that it applies really to state agencies.

But I think it's a starting point and it's important. It's common sense. So with that I'll ask for an aye vote and I thank the Chair for your indulgence.

Thank you. I apologize for forgetting about the close. Thought you done so well we didn't even need it. Let's call the roll.

[ROLL CALL]

That has five votes. We'll leave it on call for absent Members. Thank you, Senator. No Senators in the room.

[Roll Call]

Well, actually, ladies and gentlemen, we are going to take a break for one hour. We'll be back at 1pM thank you.

Sure. The order you want to do 11 or 720.

Whatever you would like is fine with me. All right, dealer's Choice.

If you're doing 720, I'll step back.

Since you're here, we'll start with SB 11. All right, we'll start with the witness that followed me up.

Okay. All right. Thank you so much. I'm here to present two bills, but we'll start with SB 11. I'd like to begin by accepting this Committee's amendments. Thank you very much for your help and assistance. The AI Abuse Protection Act establishes a legal framework for regulating artificial intelligence, voice, image, and video cloning technology.

The rise of artificial intelligence presents a lot of opportunity for innovation. Mostly is a really good thing. This innovation crosses various industries. The technologies are a powerful tool, we know that and they are obviously working towards the continued advancement of our great state.

However, there is a lack of comprehensive legal framework for addressing deep fakes and non consensual images and videos that are deeply troubling. This leaves individuals vulnerable to various forms of exploitation, identity theft, scams, misinformation, and misrepresentation of character. Unfortunately, the technology can sometimes be disproportionately used for these negative uses.

And often that is impacting women and girls though not exclusively. But the creation of sexually explicit photos and videos has taken over the Internet. It's, there are many well known cases like the Taylor Swift case which gained over 45 million views on social media in just 17 hours.

Similar instance occurred to singer songwriter, Billie Eilish. Of deepfake videos found on the Internet, over 95% are sexually explicit and feature women who did not consent to the creation. While some deep fakes target public figures, easily accessible AI software allows users to create non consensual content featuring anyone.

SB 11 addresses the misuse of artificial intelligence technology by establishing guardrails. These guardrails protect consumers from harm by one, clarifying the existing definition of likeness to include AI content. Two, requiring a consumer warning on AI software. Three, establishing violations for the misuse of AI technology. And four, preventing AI generated evidence tampering in our courtrooms.

This bill provides a balanced approach of artificial intelligence regulation while still providing remedies for victims of abuse and allowing for the industry to creatively move forward in our economy today. With me, I have a really wonderful witness who has a ton of experience in this space whose name is very difficult to say.

So I'm gonna let him introduce himself so I don't make a mistake, but very happy to have him with us today. Tell them your name and title.

Thanks for sheet. Thank you.

Honorable Vice Chair and Member, thank you for the privilege of presenting here. My name is Jai Jaisimha and I'm testifying in support of SB 11. I'm the cofounder of the Transparency Coalition. We're an independent nonprofit which advocates for increased transparency and accountability in generative AI. SB11 is not about broad, sweeping regulation of AI technology itself.

We heard about other regulations that were going in that space is a very specific bill that establishes clear definitions and rules for how high quality digital voice and image replicas can be created by AI. There's clearly evidence of consumer harm, as the Senator shared, and the bill does several things to fix that or address that issue.

High quality voice and image replica should be treated as the actual voice and likeness of the system. So that's the first thing. The second is that it expands the rules around false impersonation to include these digital replicas. It requires a consumer warning, as the Senator stated, against the misuse of these tools.

And finally, it analyzes the impact of AI based content on the rules of evidence. These are very common sense measures, in our opinion, and promote greater safety, transparency, and trust around digital replicas for all Californians. California must not be deterred from its responsibility to protect its citizens.

And in the rapidly evolving landscape of California, leadership is more important than ever. While people are taking steps to do more careful development of models, they need to be even more careful. And hopefully the penalties and the enforcement provisions in this bill will help them make better decisions.

SB 11 sends a clear message that thoughtful, specific, targeted regulation is not the enemy of innovation, but its essential partner. As a former entrepreneur, I concur and I urge the Committee to stand up for California and vote aye for SB11.

Thank you. Any other speakers or speak. Come to the microphone to speak in support, please.

Robert Singleton with Chamber of Progress, in support.

Any other speakers in support. Any speakers in opposition, please come to the microphone or come to the table if you'd like to speak as a key witness.

Thanks.

It won't be awkward at all.

At all. I'll just talk to you, Senator. Oh, gosh. Jesus. So I left my testimony in this building somewhere, so this is off the cuff, so bear with me.

Oh my God. Really like sitting somewhere in this building?

Yep, it's downstairs somewhere.

What did you say?

I left my testimony that I printed downstairs somewhere. So we're just gonna play along. Madam Vice Chair, Members of the Committee, Laura Bennett, on behalf of the California Chamber of Commerce in oppose unless amended position We very much appreciate the Senators recent amendments to the bill. In fact, we appreciate critically the change from misuse to unlawful.

We think that goes a long way to remove some of our concerns. Unfortunately, we do have some remaining concerns, namely as it relates to the penalty provision in the bill. Right now it's $25,000 per day for not providing a consumer warning.

We would appreciate that being lowered a bit to about 5,000 dollars which we think is in line with other consumer warnings such as Prop 65 which gives a person a warning about the chemicals that could cause birth defects. It's not about the chemicals causing, it's not about the warning causing the birth defect.

It's the warning is telling someone about the conduct of the chemicals. We think that's more in line in this bill. We are telling users other technology that they could could be lawful for criminal civil liability.

In addition, the new wording of the language in the bill today says unlawful use of this technology to depict another person without prior consent may result in civil or criminal liability for the user. We do appreciate the addition of consent. We think that that actually happens what we're telling people is unlawful.

Unfortunately, we think that should create a digital replica without consent. We think that is in line with the actual prohibition in section A. We have asked previously to do more amendments related to specifically marketed and designed t create a digital replica. We understand the Senator has not is not inclined to accept that amendment today.

In addition to that, we have asked to address the B2B and employee use. Again, we think the consent provision also addition may address that concern. But we are reviewing those amendments. They just went to print, I think, on Saturday. So we're still looking at those.

Tim Lynch on behalf of the Entertainment Software Association which is the video game industry and we'd like to thank the office and this Committee for their diligence on the bill. I think our oppose unless amended letter crossed the amendments in space in real time.

So we do think the amendments, our letter was largely focused on the business use cases and talking about video game creators and the business to business uses there. The move to consumers I think goes a long way in fixing the bill. We do, we may, we're still analyzing the amendments.

We may share some of the same concerns about the wording on the warning and then also the placement of the warning and how that might exist in a real time video game setting. But the amendments do seem to go in the right direction. We appreciate the author's diligence.

Do we have any speakers in opposition if you could come to the microphone, please.

Kim Stone, Stone Advocacy on behalf of the California District Attorneys Association in support. Sorry for being late.

Sorry. Also late, Connor Gusman, on behalf of the Teamsters California, in support. Thank you.

Robert Boykin with TechNet. Align our comments with the CalChamber. Thank you.

Good afternoon, Chair and Members. Naomi Padron, on behalf of the Computer and Communications Industry Association, we would align our comments with the CalChamber. Thank you.

Thank you.

Apologies also Connor Gussman, on behalf of SAG-AFTRA and support as well. Thank you.

Mix. That's good. Very diverse. Well, we don't have a quorum, so we will. Do you want to. Certainly. Go right ahead.

So the AI world has been so fascinating to me and it's so impressive what it can do. But I will tell you, scrolling through social media, I have to double take. Sometimes when you see videos and you're like, okay, is somebody saying something crazy or is this really happening in today's world? It's very realistic in either direction.

And now, you know, being a public figure sounds, I guess, kind of arrogant, but I am. It scares me of what somebody can do to use AI, to use my image and likeness and things like that. My only concern is to what you mentioned of the $25,000 figure that I do think that's extremely high.

But I've gotten to know you and. I know that you work really well to find common sense solutions. So I will be supporting the bill today. But that is my only concern is that number. So I ask that you work with opposition on that.

But thank you for bringing this very important bill forward and let's protect our image and likeness. So thank you very much Senator.

Any other comments from the Committee? Ms. Pellerin, do you have any comments? You have no comments? No. I was going to say the same as my colleague. The fine to me seems a little stiff, so I hope there'll be further amendments so we could bring everybody together on this.

I appreciate that. Such the end of the week, we're all so tired. It's so bad. It's so bad.

So I've been informed that we can take a vote. It's obviously we need a motion, a second. We don't have a quorum, but it will still be recorded as a vote. No. What do we say?

Bear with us. We've been at it for a little while. Time for a break. Time for a break. You're here. You made it. RBK's probably texting me right now.

Yeah, yeah. We need a Motion in seconds.

We have a motion.

Just. Do you need to establish your quorum first? Oh, you did? Oh we did earlier today. Yeah, that's right. Don't let me help. Sorry. Let me back. Sorry. Yes. Close. If only we were actually closing for the week. And done. We're obviously all totally done.

I just honestly, respectfully, thank you so much for taking this up. Thanks for hanging in there this week. We've been working with them. I know everyone that walks in this room tells you respectful opposition. This is actually super respectful.

They actually haven't even really had a chance to review the full amendments that have happened, so give them a chance to do that. We're open. We're open. We'll keep talking to them. And I appreciate your support. Thank you and I ask for an aye vote.

Very good. Let's vote.

[ROLL CALL]

I do. Thank you.

All right, well, hold that. Okay. Senator, you have one more.

And is SB720.

Right. SB720. I'm sure my witnesses will make their way forward. This is the Safer Streets Act, which I am very honored to present to you today. Thank the committee chair and staff for their assistance on this bill. Very helpful.

SB720 allows cities to opt into a new red light camera system for high collision zones, providing a vital tool to reduce traffic fatalities. In 2021, over 4,000 Californians lost their lives to traffic collisions, with more than a quarter of those deaths resulting from drivers who ran red lights.

The outdated and expensive structure of our current red light camera programs have proven themselves to be incredibly ineffective, overly punitive, and overly targeted at factors that have nothing to do with traffic safety. The disdain for our current red light program is, I believe, the thing that unifies us all in the legislature.

Currently, unlike other states, California holds drivers responsible for violations by requiring a photo of the person's face as they run through the red light. Driver cannot be identified, then the violation notice is sent to the vehicle's registered owner. And then, of course, the finger pointing begins.

As we all know, this controversial facial recognition is really outdated and unnecessary. And California, the home of innovation, needs to get on board with a better system.

Right now, if identified, which is quite the qualifier, violators are subject to a criminal penalty which includes a fine upwards of 500 dollars, points on their driver's license, and the very popular traffic court. As we know, insurance rates are also negatively impacted by red light violations.

Those high fees and penalties associated with the current outdated red light traffic enforcement system create a pretty big financial burden and long-term effect on families dealing with poverty. SB720 provides an updated framework so that California can join the rest of the world in the year 2025.

It shifts the program from driver liability to owner liability, removing any need for facial ID programs. Think about when you drive across the toll bridges and the camera takes a photo of your license plate. It doesn't care who the driver is, only cares that you have crossed the bridge.

Changes to the violation in terms of criminal penalty to a civil penalty. There's no reason for a red light violation to be criminal and thereby have all of the components to a criminal violation. Civil penalty suffices. This eases insurance burdens because it no longer deals with points on your driver's record nor your insurance.

It implements a tiered violation system for violations and directs funding back towards safe multimodal transportation programs in your communities. Those of us who have served on the local level we know how hard it is to find those rare and precious dollars for bike lanes and overcrossings, and pedestrian walkways.

This is a real and true funding source for all of those incredible programs that do a lot more for traffic safety than an outdated red light camera system. Local government may choose to opt into this program. It's not mandatory, but here's what they have to do in order to be eligible.

They have to develop guidelines for screening and issuing the violations. They have to develop guidelines for selecting the location to install these systems. And those locations have to be based on collision analysis where the accidents are actually happening, not where we want to police communities, but rather where accidents are actually happening.

And they must submit an annual report on the data from the automated traffic enforcement system. SB720 ensures that cities have tools needed to create safer streets, making this a critical step towards reducing traffic collision in California and ensuring that our communities are safer.

You will hear from some of the witnesses today how this system has reduced red light running and reduced collisions in other parts of the country. The bill in its implementation should do the following: It should make us safer. It should decriminalize traffic violations.

It should reduce fines, fees, and insurance burdens, and most importantly, it should save the lives of our constituents who are driving around the Golden State. With me today is Damian Kevitt from Streets Are for Everyone, and Deb Banks from Sacramento Area Bike Advocates respectfully ask for your aye vote.

Thank you. Please proceed. Whichever, whoever is first.

All right, I'll go first. Good afternoon, everybody. My name is Debra Banks. I'm the Executive Director of Sacramento Area Bicycle Advocates, also known as SABA. We are a co-sponsor of SB720. I've been in my position for six years, and I came to it as a victim of vehicular violence.

While riding my bike, a drunk driver hit me from behind at 45-plus miles per hour. I'm really lucky to be here with y'all. And here we are. SABA advocates for safe streets for all users. We believe that if they're safe for pedestrians and cyclists, then they're safer for drivers too. But we have work to do.

Sacramento is one of the most dangerous cities for road safety in California and ranks pretty high across the nation. Red light running is a road safety issue for drivers, pedestrians, cyclists, and, in particular, children. Across the U.S., fatal crashes caused by drivers running red lights have reached a 10-year high.

Sacramento used to have a red light program. It was managed by Sac County. Data collected by the city reports that between 2020 and 2023,4,300 crashes were due to a broadside or that T-bone type of a crash. 24% of those were fatal or caused injury, serious injury, and 23% of those broadside crashes included a cyclist.

In 2023, there were 1,300 broadside crashes alone. And Sacramento County dropped our red light camera program in early 2024. Data highlights the danger of intersections, as Senator Ashby just said, 4,000 annual collisions nationwide. In Sacramento, more than half of those annual collisions are at traffic signals. According to AAA, three people die in red light running crashes nationwide.

Yesterday, I witnessed three drivers run a red light in the five miles that I commute on my way home from my office. They're not just statistics. They're preventable tragedies that demand immediate action to fix our outdated traffic safety systems. Every community in California deserves access to proven safety tools that protect pedestrians, cyclists, and drivers alike.

Thank you. Next speaker, please.