Assembly Standing Committee on Privacy and Consumer Protection

Welcome to the Assembly Privacy and Consumer Protection Committee hearing. I am not your normal chair. Our current chair is presenting a Bill. I believe in judiciary and we do not have a quorum, so we will start as a Subcommitee to effectively manage our time. Today, we will be limiting testimony to two witnesses for support and two witnesses in opposition on each Bill. Each witness will be allowed two minutes to present their testimony.

After the support witnesses conclude their testimony, the Committee will call up additional supporters. No further testimony will be permitted. Additional supporters may only state their name, affiliation and position for the record. The same process will be followed after the opposition witnesses conclude their testimony. At that time, additional opposition will be called. They may only state their name, affiliation and position for the record. For this hearing and until further notice, the Committee will be returning to in person testimony only as reflected in the file. Notice.

I'd like to note that we are accepting written testimony through the position letter portal on the Committee's website. Let's see. I'd also like to welcome some Member Chen, who's not here yet, who will be filling in for some Member Hoover for today. Privacy and Consumer Protection Committee hearing. We have eight bills on the agenda. AB 2461 by Assembly Member Mathis was pulled from the agenda yesterday. We will welcome our first author.

Well, we're going to welcome our first author before we establish quorum, because we have a Subcommitee right now, so we don't need to call the role. Okay, we will first hear from. Let's see. Assembly Member Weber. AB 1791. You may begin.

Thank you. Chair and Member of the Committee, I am here to present AB 1791. I will be accepting the Committee amendments. With over 95 million images uploaded to Instagram every day, it is important to ensure that users can determine where an image originated and if it is AI generated.

But this can't happen if social media platforms remove any data from an image, that would tell users digital information about the media, such as the system it was created on or when it was created, AB 1791 would prohibit social media platforms from deleting certain digital providence data from digital content. This Bill defines both personal providence data and system providence data and requires that social media companies keep system providence data but deletes personal data.

In the case that the system and personal providence data could not be separated, this Bill requires a social media company delete both, but attach a label to disclosing the system's providence data was deleted. Finally, this Bill is enforced by a private right of action. At this time, we acknowledge the concerns with the PRA. After discussions with the Committee, it was determined that the Bill needed some sort of enforcement.

Unfortunately, the fiscal deadline had passed, so we had to add an enforcement mechanism which did not cost the state money. I know the PRA is not welcome and I want to express my commitment to continuing to work with the Committee and opposition to determine the best course of action for enforcement. I do want to thank Committee and stakeholders for the thoughtful process and look forward to continuing to work on this Bill as it continues to move forward. I respectfully ask for your aye vote. Thank you.

Great. Thank you. I'm glad the Committee Members all got to hear that. Do you have any primary witnesses in support? Okay, any primary witnesses in opposition? Thank you. You have two minutes.

Thank you Mister chair Dylan Hoffman on behalf of Technet and unfortunately, we are respectfully opposed to AB 1791. As the author noted, we had previously suggested amendments to this Bill without taking a formal position, and we were having productive conversations with the author's office. However, the recent amendments significantly expand the Bill. As mentioned, it does include a new private right of action which greatly exacerbates the challenges for platforms to comply with this Bill.

And I think we appreciate the commitment and the recognition that this enforcement mechanism isn't necessarily appropriate to this type of policy. But we have to analyze this Bill as it currently exists in print, which leads us to our position. And so, as we've previously stated in our concerns letter, we agree with the intent to create greater trust in user content, and we even agreed that platforms should not remove what this Bill now calls system provenance data.

However, the new changes now require platforms to remove all personal provenance data, defined as any metadata that includes personal information. First, some of that information is useful for provenance and authenticity, such as the location, the time of day, and even the creator's name, particularly if it's a photographer or a digital artist. They may want that included in that data. Second, it seems to require the removal of personal information.

Even if a user consents and indeed wants it there, it would be on the user to add that information back in somehow. The Bill then requires that if a platform can't remove the personal provenance data without removing the system provenance data, that it should remove both. Once it has removed both, it is then to re embed the system providence data with an indelible label that will remain when the content is reshared, downloaded or uploaded to another platform.

However, this Committee has previously passed AB 3211, requiring that if AI generated content and recording device manufacturers create maximally indelible watermarks, then it's unclear how a platform would go about removing that type of information. Either the platform can remove it, in which case the AI developer or camera manufacturer is liable for not creating a maximally indelible watermark, or the platform can't, in which case they're reliable. Under this Bill, the technology to redact this kind of information and re embed it does not exist yet.

And as mentioned, with the immense volume of content that is shared and uploaded on these platforms, sometimes 95 million per day, having that ability would require significant resources. And we think engineering this type of system now would detract from current investments to develop the sort of tamper evident, maximally indelible provenance technology as required by other bills such as AB 3211.

If you wouldn't mind wrapping it up.

Thanks, appreciate it, Mister chair. So unfortunately, these recent amendments do cause us to oppose the Bill at this time. We look forward to continuing the conversation with the authority should this Bill move forward, but respectfully ask for your no vote. Thank you.

Great, thank you. Any other witnesses in opposition, please say your name and affiliation.

Thank you Mister Vice Chair Members Ronak Daylami, on behalf of Cal Chamber, respectfully also opposed. Further reasons stated by my colleague. Thank you.

Thank you.

Good afternoon. Naomi Padron on behalf of the Computer and Communications Industry Association, we are respectfully opposed. And would echo the comments made by my colleague at TechNet thank you.

Any comments from the Committee Member. While he kind of figures out, I have a question. I agree with where you're going in terms of having this personal information removed, especially people, including me, just upload this stuff. I make that choice, obviously. But doesn't the personally identifiable information somewhat help to make sure it's not AI? You know what I mean?

Yeah, no, I completely agree. And we will continue to work with opposition on this. I think there were some changes with the amendments from this Committee, but we'll continue to work with the opposition to ensure that we're able to have the content removed.

That is personal that someone wouldn't want, like, I don't want you to necessarily know that I was here at this particular time when I uploaded this photo, but still allow for a user or someone who's watching it or looking at it to understand that this is AI versus a non AI generated image.

All right, thank you. Mister Ward, did you have any.

Thank you. Yeah, and as I was getting ready for Committee, I know this is one that I was trying to draw closer attention to. Apologies for walking in the middle of your presentation and trying to kind of reconcile here.

I had equal, I had similar struggles as well with the previous Bill that we talked about in this Committee when it came to providence data being kind of a newer thing, newer definition, and something about how do we get around regulating that and to, to what appropriate extent should we be.

So on one hand, the Bill, you would want to see providence data incorporated, where GenAi is used to alter an image so you can kind of track to its original source, but we want them also to have removed that for maybe original content.

So it's removing the personal data. So some people may not, like I just said, you may upload an image and you don't want someone to be able to determine that you were in San Francisco at a particular time, but you do want to be able to determine that that is a real image versus an AI generated image.

And so walking that line of maintaining someone's privacy if they don't want that image out there, but also still making sure that we're getting the information that this is real data versus fake.

But if I. So in some cases, we would want that. You propose we would want that attached to the file, and in other cases we don't. But why would that be the responsibility of the social media company as opposed to the person uploading it?

And I think that's what I just specifically was stating, that we'll continue to work with opposition to see how we can, because some people may want that, as I think Assemblymember Patterson or opposition had stated, some people may want that personal data out there. And so we will continue to work with opposition to figure out a way in which we can remove the data if people don't want it, but allow for it to stay on there if they do want it.

I mean, I think it's easier to say that if somebody doesn't want it up there, they don't. They can probably upload a file. I'm not, you know, enough tech savvy to know that they could upload a file that doesn't have that data included. I'm sure there's a mechanism.

Well, I'm sure most people aren't that tech savvy. I know I'm not. And I upload things to social media all the time, and I have no problems with someone seeing that it's a real image, but I don't necessarily know if I want them to know that I took it at a particular place at a particular time. So I'm not sure if it's that easy and recognizing right now that the technology for these tech companies may not necessarily be there. And so we are in conversations for delayed implementation and things like that.

But if you're the company, you're trying to regulate what is being uploaded and whether or not it should or should not, then have that data attached to it for the, you tell me, is it millions, billions, trillions of files that are uploaded there? How can you discern whether or not file a versus file b, which was uploaded by two different users, is subject to the requirement to have that attachment?

Right. And that's why with this particular Bill, it just removes the personal data. So like the name, the location, and those kind of things not allow for that flexibility because it is complicated.

But you want that for anything that uses GenAi?

So what we want is to know that this was actually real versus AI. Right. But you don't necessarily have to have someone's name and specific location to know that it is real versus an AI.

Okay, and how would that then be than different for the intents of Miss Wicks Bill that we heard before?

Oh, so great question. So with AB 30 to 11, that one requires labeling. So what we're talking about is what's happening on the back end. Her Bill is what the actual. And if she were here, she could talk about it better, but what the actual user would be able to see on the front end. So we're talking about the back end data that you and I don't see. What hers is talking about is an actual label that you or I would see. So it is, it is different. And they could work together or they could work separate, but it's not the same.

Okay. Okay. Let me reconcile all this. I know that you would continue to work on these issues very strongly as well. It's obviously an area that we're both growing and understanding, but trying to get right for a lot of benefit for consumers. And certainly, I think, to your bill's point, for the privacy issues that you're trying to ensure here as well. So I'll study a little bit closer. I know we'll show as a Subcommitee, it'll give me a little bit of time to kind of think through those. Thank you.

Sounds good. Thank you.

Thank you. Any other questions? All right, if you want to give a closing statement, that'd be great.

Yes. Well, I really want to thank once again, the sub, the staff on this Committee for working through this Bill. It is a very complicated issue because it's a jew. And you know, we definitely commit to continuing to work with the opposition to really iron out some of these concerns that they have. And you know, and I still am working through with some of this Bill. But as this Bill moves to allow this Bill to continue to move forward, I respectfully ask for an aye vote.

Thank you. We'll do all of that when there are is a quorum here. So thank you very much. All right, Mister, Mister Ting. We have AB 1814.

Thank you.

Thank you, Mister chair. First, let me just thank the chair and the Committee for working with me on this Bill. We're happily accepting the amendments. This Bill has to do with facial recognition. As of right now, there is no regulation of any facial recognition technology in regards to how it's used for law enforcement. This Bill will prevent law enforcement from being able to arrest, search or get a warrant when the sole basis for that action is a facial recognition match.

So again, we would be trying to restrict it so that law enforcement would have to have other evidence besides the facial recognition match to be able to get people arrest, search or get a warrant. In previous years, I've been legislation on facial recognition software. I did a Bill where we had a moratorium for a number of years. That moratorium has now expired. So because that moratorium expired, there's absolutely no regulation on this issue.

Last year, I tried to do a much more extensive Bill regulating facial recognition technology, but there was a number of concerns with that. So we decided to only focus on the false matches to ensure that the match itself could not result in an arrest, a warrant or a search. So with that, I do have Jake Parker, who's senior Director of government relations at the Security Industry Association, as a witness.

Thank you, Assembly Member, Madam Chair, Members of the Committee. So again, Jake Parker with the Security Industry Association, SIA, represents more than 180 companies headquartered here in California that provide safety and security products. And among our Members are the leading providers of facial recognition software for investigative tools used by law enforcement agencies in our nation.

So this technology is fielded by some California agencies for more than 18 years and is used daily to assist in identifying and capturing our most violent criminals and bringing justice and closure for victims. This software, which compares photos for similarity, is the only way, in some instances, to generate leads regarding the identity of a key witness, suspect or victim. Without it, many cases would go unsolved. In particular, today's best tools for fighting human trafficking and child sexual exploitation simply would not exist without facial recognition technology.

Now, same as other sources of identity, leads from an image investigation, like making public announcements, soliciting anonymous tips, investigators must find and confirm further independent information in order to make a positive identification, and if they are a suspect, to establish probable cause.This Bill addresses I think what is the primary public concern about use of the technology is that in that inaccurate results could cause misidentification of suspects or even wrongful arrest.

Unfortunately, wrongful arrest remains an everyday occurrence in the United States and there's many different causes for this. This while there are just seven reported cases where use of this technology is alleged to have caused such arrest, we believe this should never be the case. It should never be a contributing factor, because the investigative process, in the investigative process, it does not confirm an identity. So this Bill, we believe, will help ensure leads are always cooperated with actual evidence before taking action to address this concern.

And in fact, all stakeholders that support having uniform rules for law enforcement use the technology, including SIA, agree on this baseline parameter. There's a clear need to bolster public trust that technology is being leveraged accurately, lawfully, and an effective way that's also limited and non discriminatory in a way that benefits our communities. So we believe AB 1814 will help bolster this trust and for that region, urge you to support this Bill in its current form.

Thank you very much. Any other witnesses in support?

Mister Chairmember Jolena Voorhis, on behalf of the League of California cities in support.

Becca Kramer Mauter on behalf of ACLU California Action and respectful opposition, we appreciate the authors and the Committee's understandings of the harms of face surveillance and your desire to protect Californians from the compounding harms of false matches leading to wrongful arrests. However, AB 1814 won't provide that protection. A nearly identical policy has been present in almost every known case of wrongful arrests involving this technology, and it has routinely failed to prevent the life derailing, misuse and harmful mistakes caused by this technology.

And because the bar in AB 1814 is so Low, the enforcement mechanism being added today will not have the desired effect. Indeed, it would not have provided redress for most of the known victims of wrongful arrests based on false facial recognition matches had it been available in those cases.

As we have seen elsewhere, policies prohibiting facial recognition as the sole basis for probable cause incentivize police to launder potentially flawed facial recognition results into other parts of the investigation, including photo lineups and witness interviews, so that facial recognition is not the sole basis for probable cause. In addition to tainting the investigation, this would negate any redress for someone wrongfully arrested due to face surveillance, if they even ever found out face surveillance was used in their case.

Additionally, well intentioned policies like those in AB 1814 don't just not work, they also can cause real harm by signaling that there is an acceptable way to use facial recognition.

Facial recognition taints police investigations by layering three psychological biases that, once introduced, cannot be undone the first, automation bias, or the very common assumption that the tech is right despite any evidence to the contrary second, confirmation bias, where officers are predisposed to find evidence that confirms the false match and third, racial bias, which extends beyond the false match rate to the use of this system with police mugshot databases that are essentially a record of racially discriminatory policies like the war on drugs.

For these and other reasons, we must respectfully oppose AB 1814.

Nathan Sherd for the Electronic Frontier foundation the Electronic Frontier foundation respectfully opposes AB 1814. By setting up guardrails that have already proven ineffective, AB 1814 would codify approval for law enforcement to use face direction technology to identify and track California residents throughout the state, and it fails to provide any measure of accountability for officers that ignore the inadequate guidelines that it provides.

As the support states, there are seven cases that we know of where folks have been wrongfully identified by the technology, and in five of those warnings, like those mentioned in ABH 1514, were already in place when police ignored these warnings, as we've seen officers do repeatedly across the country. There is no mechanism through which Californians would learn of that noncompliance, nor does it provide a pathway to enforce the rule.

Last year, Robert Williams, who was wrongfully arrested in front of his children, his wife and his neighbors, told this Legislature of how the only way that he was notified that face direction technology was used in his case was because one of the officials interviewing him slipped up. Those warnings were in place. Face direction technology, as my colleague has said, taints police investigations in many ways, including automation bias, a term that describes how people and police trust technology even when it's wrong and contradicts observable facts.

That was the case in the incident of Michael Oliver, who is 26 and verifiably at work. When a teacher had their phone taken from them and damaged, that phone had captured image of the assailant, who had no tattoos, unlike Michael Oliver, who had face and neck tattoos and again, was verifiably at work. That did not stop him.

And warnings like those offered by AB 1814 did not stop police from arresting him and ultimately him losing his job and his ability that it provided to support his son and his family and his home. Once again, warnings like AB 141814 puts in place are seemingly ignored. Rather than aligning with the civil rights community and national consensus, AB 1814 grants law enforcement agencies the authority to use face directness technology to identify and track people across the state.

And it creates ineffective guardrails and requires no accountability if law enforcement enforcement abuses them. For these reasons, we oppose AB 1814 and respectfully urge the committees. No vote. Thank you.

Thank you so much. Seeing no further opposition in the room, I'll bring it back to the dais. Mister Ward, great.

Thank you. Any additional witnesses in opposition to this?

Thank you, Madam Chair. Thank you, Madam Chair. Due to their unavailability to be here, I was asked to state for the record the opposition, the following organizations secure Justice, Oakland, privacy, San Francisco public defenders, racial justice Committee and tech equity. Thank you.

Thank you for bringing forward this Bill. When I was in local government, we've had a lot of conversations around smart streetlights and the use of facial recognition technology. The need, at least for our city to be able to establish a citizens Commission to understand when technology should be used or not be used, and then, importantly, how the data can be stored. Who has access to that data, say, within a Police Department? And under which conditions can somebody access the data?

For example, you wouldn't want somebody, an errant officer deciding to go after somebody's ex girlfriend or something. Right. So definitely something that really interests me because I think it's important that we get these issues right. But what I have trouble understanding, I guess the opposition can explain a little further. Without this Bill, the way things are today, a local law enforcement agency can use facial recognition technology, correct?

Correct. They can, although there's legislative uncertainty, and so they may choose not to adopt it because it's not been affirmatively approved by the Legislature at this point.

Well, but it's also not affirmatively denied. And we can have that conversation about, again, I think we should around guardrails and about, you know, the use and utility and when to use it or not use it and all the protections that need to go around that. But all this Bill does is says that you can't have facial recognition technology information as the sole basis for making an arrest. Right. So.

And we do know that there has been information provided that has helped lead to an arrest of a crime that actually did happen that otherwise may not, somebody may not have been pointed to in a certain direction. So if it is supplementary and it's informative, but not the sole basis, I agree that there is huge disparity and there have been unlawful and wrongful arrests just purely based on biometric technologies, why couldn't this be helpful towards maybe the same end that you're achieving?

So there certainly are. There may be incidents where it can be helpful in identifying a suspect. There may be a small section of the population that helps.

It will also have a chilling effect on the engagement of First Amendment activity for large swaths of the community and the impact there and the requirement to make sure that the disparate impacts that already exist in technology and in addition to the ways that it would chill essential freedoms for large segments of the population, are things that we really need to take importantly into account.

And again, the guidelines that are offered here, without accountability measures, without transparency measures, would really just codify the ability to use the technology. And we've seen that the warnings and the guidelines that it does offer have been ineffective in preventing arrest and trauma to families and individuals as a result of the wrongful associations.

And if I can add on to that, if I may. If I may. So I also want. I know this analysis came out not super long ago, so I'm gonna give everyone a benefit of doubt. They didn't have time to review it, but if I'll refer everyone to page seven of the Committee analysis, and I believe I was not here. But Mister Ting is accepting Committee amendments. One of the Committee's concerns was enforcement.

And he has taken additional amendments to clarify that anything that violates this would constitute false arrest with specific penalties. So I keep hearing the opposition say that there is no enforcement or accountability and that is no longer true. I want that to be clear. Thank you. I appreciate that. And I apologize if you missed.

May I respond to you as well? Wanting to walk through a little bit more. And I would encourage you. I'm sorry that it got to you all so late that one of the papers that I just passed out is hot off the press. A blog today that came out today, walking through how these policies not only fail to prevent wrongful arrests, but actually cause harm in and of themselves.

I understand your opposition, but he keeps saying there's no enforcement. So I just wanted to make sure we were clear on what the Bill currently does. But it's Mister Ward's time.

I think you came in part way while I was going through, and I did explain how as much as we appreciate the effort to put in enforceability, because the bar of what is required by AB 1814 so Low, even in the false arrest.

And one of the ways that it causes harm, in addition to the three kind of psychological biases that is built in and amplified by facial recognition, and it's impossible to separate out from facial recognition, and I can go into those more if you would like, is the way that it is used subsequently to get past the sole basis requirement.

And so one of the things that we've frequently seen in the cases where these false matches have led to wrongful arrests, even in instances where a policy like this one is in place, is that then what happens to get around the sole basis is that law enforcement takes this false match.

That isn't saying this is the suspect, it's saying this is a doppelganger of the suspect, puts it into the witness, the photo witness lineup, and in doing so, is going around the best practices and norms that have been established for making sure that photo witness lineups, pre facial recognition, were accurate and unbiased, puts it in. And so then it's no longer is this person the suspect that you saw in this lineup, it is the doppelganger of the suspect in this lineup.

And now they're going to become the suspect. And even worse, we oftentimes see these biases, like automation bias, come in and then mess with the photo identification. So my understanding from lawyers who work in these cases is best practice, is that the officer administering the photo lineup does not know if they're like, who in the lineup is. The suspect is kind of removed from it, so they can't bias it.

What we have seen is that officers who are administering this lineup have been told themselves, and even then, have told the witness, the witness that both before they identify someone, that there was a facial recognition match. And then we get into the automation bias, like what my colleague was talking about, where even when the suspect in the photos lacks any tattoos on their face, and the doppelganger does have tattoos, and it's clearly a mismatch. Everyone assumed, even experts assume, that the tech can't get it wrong.

The tech is right, I must be wrong. Therefore, what is obviously a false match is correct. And that starts to come in and then taints the witness ID. Because suddenly the witness, having heard that there's a match from facial recognition, their automation bias is someone in this lineup is for sure the suspect. Even if I don't actually see the person. Even if I wouldn't have identified anyone in this lineup as the suspect afterwards.

But that way, they've now met the requirements of AB 1814 by having done something in addition to the facial recognition.

Match, if I can. The person, though that is, say, participating or is trying to make that identification, they're not so far biased, is that they're given a copy of the facial. The image that was taken and used for facial recognition technology. Right? They're not. They're saying a match is made. Okay, fine. One of these six people has a match made. Maybe that's information that somehow is going to buy someone's thought process, but they're still having to make an independent judgment of whom they saw.

But suddenly it's no longer an independent judgment of is the person you saw on this page. But the technology, who everyone trusts to make smarter decisions and be more accurate than themselves, has said, the person you saw is on this page. So now tell me which of these people looks most like the person that you saw, even if you think they're different, because the technology is right, and the technology has said the suspect is.

Here, but presumably the person is the victim, is not going to have access or shouldn't have access to that image.

But it's going to show up. Because law enforcement is using the image that showed up in the facial recognition match as one of the photos in the lineup. It's not the photo from whatever surveillance camera or what else. It's the match photo that shows up in.

So you'd have two images, if we.

I think Mister Ting would like to clarify his intent.

So this is actually a very simple Bill. Let me just cut straight to the chase. The opposition here today, if they were going to answer this question, would only support a Bill that banned facial recognition. That's it. Both of them. Let me just finish. They talked, you heard the electronic Frontier foundation talk about all these issues around First Amendment rights and around a number of these issues they addressed. My previous Bill last year addressed, I believe, almost all of them, and they still opposed.

So again, this Bill, while I wish this Bill would stop wrongful arrest, I wish it would solve racial bias, I wish it would bring peace in our world. This is a very narrow Bill to do with a very particular issue around false identification and facial recognition. What the witnesses have managed to do is talk about a variety of issues that we are, that are problematic within our criminal justice system, that are not part of the Bill.

They have also talked about particular situations that I believe happened outside of California. Right. Your question, which is the seminal question, is right now, law enforcement has every ability to do whatever they want with facial recognition software and technology. And last I checked, unless the witnesses want to disagree with me, I don't believe law enforcement needs the Legislature's permission to go do something. So again, this would provide a guardrail against it.

The Bill I had last year had a lot more guardrails, of which they opposed because it wasn't good enough, because it wasn't a ban. So again, this is just one guardrails I would have last year. I had many, many more, a much more complex set of regulations. I think a much thorough Bill.

This was the one area that I think there are a number of people had, had concerns about a false match, making sure that the false match did not result in an arrest, a search or a warrant. And so that's why. So this Bill is a very narrow Bill.

Okay, I'll just summarize. Thank you, Mister Ting, that I'll just summarize my comments. I do think, you know, on his face today, that I see that this is meant to be more helpful towards a better and correct outcome. Right. It's a piece of information, but it should not be the sole piece of information and that would lead to an arrest and subsequent activity. I don't.

But to the author, I don't disagree that there aren't very serious issues that need to be worked out in this General space here. That's not what this Bill does. And this could be a vehicle for continued conversation as we're seeing, for example, through Committee amendments to try to make sure that there's some balance here. And thank you for accepting those. So I'll be very interested to see how more, whether through this Bill or through any Bill in this area, that we're continuing to perfect this option. But I can see this being helpful in my view, towards a just outcome.

Thank you. I'm going to pause for 1 second to establish quorum. Let's call the roll. [Roll Call] We have a quorum. Thank you, Mister Lowenthal.

First of all, I want to commend the author for bringing this forward and, you know, attempting to chisel away at you, you know, this great gray area that we have right now and try to, you know, set expectations amongst, you know, the citizens of California on what they have, what they don't have. And I also want to commend the opposition for being the voice of, you know, protecting us all in the cases where things can be rogue and utilized against us. And I really appreciate that.

I truly do. I do have a question for the opposition where you lamented that there is no accountability for abuse. Gentlemen, I forgot your name.

Nathan, Mister Sheard, about the lack of accountability for abuse, specifically with this technology. And I wanted to understand, do we have specific accountabilities for abuse in other technologies that are utilized in law enforcement? Does that exist anywhere in our legal codes? For example, video stings, use of informants, things like that? Do we have specific legislative wording around abuse for that specific technology?

Nathan Sheard.

I can't speak to specific language and exist in reference to any specific technology. What I can say is that without the ability to know that face recognition played a role in their arrest and accusation, and without the ability to have something, a private right of action or some other way to actually be empowered to enforce when the guidelines that are provided here are misused, are not followed, as we've seen happen repeatedly, there is, we lack an ability for an accused, for a person who's been accused to have real action in that space.

And I'll again refer everyone to page seven of the analysis that has the new amendments that have a civil action for people who are wronged.

Thank you, Madam Chair. Okay, so I appreciate that response. And then I also had a question for the author. This is clarification, forgive me that I don't have my Ipad that has any battery juice in it right now. So I don't have the Bill imprint in front of me. How does this. Excuse me, jurisdictionally, is this only for California State and local law enforcement agencies?

And what happens with federal law enforcement or Customs and Border protection that's utilized in the State of California because of their widespread use of facial recognition technology and cross jurisdictional law enforcement?

I think it's a great question. My understanding is that this law would only apply to law enforcement within the State of California. That state or local, it would not apply to federal.

Okay. Because I know it's interesting, it just, you know, if you're utilizing, for example, global entry, it now uses exclusively facial recognition as a mechanism, you know, to enter the country. Right. And of course they are cross checking that with other information, presumably with law enforcement databases as people are trying to enter the country. So I would want to understand about that nexus and would hope you can take this question to heart as you're, you know, going forward on this.

Appreciate it.

No, I appreciate that. Obviously we have legal limitations on how we bind the Federal Government, but I would think you would need more than, if there's a warrant out for someone's arrest using multiple pieces of evidence, then this would not be the sole piece of evidence for such arrest, even if they used it to find the person.

That's correct.

So that would be my understanding of the Bill is written. Yeah. Mister Patterson.

Great, thank you. So just I think first of all, that any wrongful conviction, you know, using facial recognition technology or not, you know, should be penalized. And you know, sometimes accidents happen, but should never cause a person's freedom, you know, to go away. But just kind of dig into the first of all, great line of questioning I thought on, you know, because we don't want to, we're already banning clear, so don't want to ban, you know, global entry also.

But when I read the analysis, it says prohibits a law enforcement agency or peace officer from using a facial recognition technology match as a sole basis for an arrest search or as an affidavit for a warrant. So just to clarify for me, as far as you know, there is no bracket at all on law enforcement using.

There is no bracket at all. They could do whatever they want right now.

Yeah, I know. Yeah. Yeah. So I'm just that.

That was your question. They can. There is no regulation of facial recognition technology as it applies to law enforcement except for, and I forgot to. I was going to mention this with Mister Ward's questions. So there is no local preemption. So my city has banned facial recognition technology in a variety of uses, but also in terms of law enforcement, Oakland has as well. So this does not preempt any city should they choose from taking stricter measures.

Okay.

But right now, outside of those local cities that have done it, there is no regulation of facial recognition technology as it applies to law enforcement.

And with 500 cities and 58 counties, you know, I doubt very many of them have limited, you know, the use of facial recognition technology. And I just wanted to, you mentioned it multiple times, but just wanted to clarify that because I understand the opposition's concern regarding wrongful convictions regarding facial recognition technology. But it seems without your Bill, the likeliness of that occurring is actually increased if this Bill does not pass.

That would be my opinion.

And so I would say if we need to put some more brackets on it, that that could potentially be, you know, opponent sponsored Bill in the future.

I wish that were the case. I had a much more robust this.

I remember that last year.

Yeah, from last year. And they were equally respectfully opposed. I'd love it if they would just be disrespectfully in favor of it. But, you know, I got respectful opposition, not disrespectful support.

I've actually been waiting for the day for somebody to be disrespectfully opposed to some legislation. This could have been the avenue, this.

May have been the day, but today is not today.

I'm pretty sure this Committee has seen that disrespectful opposition in recent weeks

Come to Public Safety Committee.

So thank you. Seeing no other Members with questions or concerns, I again want to thank the author for taking the amendments and clarifying the way to enforce this. And, you know, I've seen your work over the years on facial recognition. I used to sit on public safety where you brought many of those bills.

And I know you and your city would like to go a lot further than this Bill, but you're also have been around this place many years and are someone who understands how to make policy and find consensus. And so you are moving something, I think, that will do something meaningful, but, you know, hopefully also may get through to the governor's desk. And I do think, having done criminal defense law myself, that we see practices that lead to arrest and incarceration on very little evidence.

You know, I don't dispute what the opposition said about photo lineups. I dealt with photo lineups in my career, and they're often misused in ways even more abusive than you mentioned today. And so it isn't as if this is the solution to all the problems in our criminal justice system. But it is doing more than we're doing today. And I think that it moves California forward. And that should always be our goal. So with that, I have an aye recommendation. Would you like to close?

Thank you. That was beautiful. I'll use it as my clothes and disrespectfully ask for an aye vote.

Item number two. AB 1814 by assemblymember Ting. The motion is do pass as amended [Roll Call] That Bill has six votes. It is out. But we will leave the roll open for absent members.

Okay. zero, we don't have a motion. Do we have a motion on the Bill? Motion. Do a second. 2nd

Thank you.

Mister Haney, you are next.

Thank you. Madam Chair. Good to see you. I will be accepting. This is AB 2741. I will be accepting the Committee's amendments today and I appreciate the thoughtful feedback on the Bill. AB 2741 is a response to the increase of rental car thefts occurring in California right now. Rental cars are being stolen at alarming rates, with rental car companies reporting that they have had an over 266% increase in vehicle thefts over the past three years.

Unfortunately, rental car companies don't have the ability to intervene in a timely fashion. Under current law. When a car is missing or stolen, tracking systems cannot be legally activated until the car has been missing for 72 hours past its return time. That means that there are three whole days where the location of the vehicle is unknown. By that time, tracking systems are allowed to be activated and what often happens is vehicles are recovered, severely damaged, or with stolen goods, mail and packages inside.

Some companies are even reporting that by the time they can locate the vehicle, the tracking system shows that the car's last known location was outside of the United States. AB 2741 is a simple, common sense solution to the increasing rates of rental car theft. It would allow rental car companies to initiate vehicle recovery 24 hours after failure to return a vehicle, instead of 72 hours.

This would mean intervention before a car has left the country, aided in a crime, broken down or broken down in a remote location. It also helps companies recover valuable assets more easily. Here to testify with me today in support of the Bill is Sharky Laguana from the American Rental Car Association.

Great, thank you. You'll have two minutes.

Thank you. First, I just want to acknowledge I have the goofiest name of anybody you're going to see today. Just want to put that out there. I'm the President of the American Car Rental Association. We, our Members are 98% of all cars that are rented in America. They include all the major companies you've heard of, Avis, Hertz Enterprise, but they also include small companies like myself. I run a company called Bandigo. We help bands go on tour.

I've been operating out of San Francisco for 20 years. We're here today to, as Assembly Member Haney pointed out, to highlight that we've seen a massive uptick in vehicle thefts. Our Members are reporting as high as 266% increase over just the past three. On a personal level, I can tell you my fleet is orders of a magnitude smaller than enterprise, Hertz, and Avison Enterprise. I personally have seen more stolen vehicles in the past year than I have seen in the past 20 years combined.

Our vehicles are being used to commit. My vehicles have been used to commit more crimes. They've been recovered with stolen items inside them. My vehicles have been. When we are finally allowed to activate GPS, it's in the vehicles in Mexico or overseas. And I want to. It's an issue for all residents because stolen vehicles are often used to commit more crimes. And car rental companies are an attractive source for people that want to commit more crimes.

Because these vehicles are not trackable to the owner like an ordinary stolen vehicle would be. They're stealing the vehicles using three methods. One is identity theft with stolen ID and credit cards. Two is counterfeit ids that are difficult for us to identify in real time. And three is they are convincing innocent people to rent the vehicles on their behalf. And they are doing this either by offering them money or sometimes they're doing this thing. I've actually personally bumped into this. They're love bombing people.

This is what I heard this is called. They convince somebody to go get the vehicle, and then they bring them back and they say, I'm going to go run to the store. And then the vehicle never comes back again and is never seen again. It is from our perspective, we're not interested in contributing to more people being incarcerated or arrested. We just want to recover our asset. All of the companies are completely aligned on this.

The sooner that we can activate the GPS, the sooner that we can recover the vehicle using our own means and get it out of the hands of somebody that potentially might be using it to commit more crimes, which does happen with significant frequency. And by doing that, we help stop crimes from being committed. We help hopefully stop more arrests from happening. We also get our asset back, which is important for us in order to be able to conduct our business.

So that's why we are moving forward with this Bill. We appreciate the amendments. We accepted them. We have no interest in consumers privacy. We aren't interested in where people are going. We're only interested in our asset when it hasn't come back.

All right, thank you very much. Any other witnesses in support?

Good afternoon. John Moffatt, on behalf of Enterprise Mobility insupport.

Thank you. Seeing no further support in the room, do we have opposition on this Bill? Seeing no opposition, we'll bring it back to the Committee. And we have a motion and a second. Any questions? Concerns? Mister Haney, you might be our fastest Bill of the day.

Camille Wagner, representing Hertz in support.

Good afternoon. Kaitlyn Johnson with Political Solutions on behalf of Avis and California Travel Association. Thank you.

I want to thank the author for taking the amendments and taking seriously the privacy of those who rent your cars, which I know is critical to you in the industry, and waiting until after they've returned the cars to turn it back on, and even after an extended contract period. So I think the Bill strikes a real balance between protecting the rights of these individuals and the privacy of those who rent them. So with that, would you like to close?

Thank you. Thanks again, Madam Chair, and to the Committee, and respectfully ask your aye vote, we.

Have a motion and a second. Let's call the roll [Roll Call]

Bill has seven votes. It is out, but we'll leave the roll up and perhaps Members. Thank you granny. Mister Rodriguez can thank you very much in time, much to Mister Low's chagrin. So this is AB 2146 with the Bill.

Thank you Madam Chair. Members, thank you for the opportunity to present AB 2146 which would ban the sale of personal flotation devices not approved by the United States Coast Guard. Before I begin, I would like to accept the Committee amendments and thank you Madam Chair, and your staff for working with my team on the amendment to clarify the type of personal flotation devices we hope to ban from the market.

I look forward to continuing to work with you and the Committee on Refining the language on this Bill. I decided to introduce this Bill to help eliminate drowning tragedies. As these are preventable incidents. Nobody should endure the pain and difficulty that come from these tragedies. According to the centers of Defeat, control and Prevention, CDC drowning is the leasing leading cause of death. For children the ages of 1 and 4 and the second leading cause of death for children between five and 14.

Individuals who suffer from a non fatal drowning can experience long term and costly health problems, including brain damage or permanent disability. While the state has taken several steps to promote water safety to prevent drowning accidents, there are consumer products such as water wings and floaties on the market that fail to keep children safe while in the water. Many experts, including the CDC and the US Consumer Product Safety Commission and the American Academy of Pediatrics, have expressed the dangers of these products.

In fact, the US Consumer Product Safety Commission has labeled these products as toys and requires these toys to be sold with an accomplished warning label. But despite the design as a toy, the labels warning parents of the dangers of these products. 66% of the parents believe floaties and water wings keep children safe in water. Therefore, this Bill proposes to ban wearable personal flotation devices designed to be worn or attached to the body that are not Coast Guard approved.

By removing these types of products from the market, we can save lives and prevent the life challenges that come from non fatal drownings with that. Thank you.

Thank you. Anyone here in support of this Bill? Do we have any opposition to this Bill? Seeing none. Bring it back to the Committee. Mister Patterson?

Yes. At the end there, you're talking about those, you know, swim toys or.

Yes.

Is your intent to ban those or.

Well, just be labeled? They are not. I guess a lot of folks think that there are a safety, but they're not. They're a toy. So we just want to ban those ones that are not sold as being a flotation device, which they're not their toys.

Okay, great. Thanks.

Thank you. Any other questions? Seeing none, one of the little known facts about me is I was a lifeguard in my earlier years and I learned that those things are actually incredibly dangerous. That they. Kids. Kids believe they can float from wearing them all the time, and then they'll go into water believing that they have the ability to float and they don't. And so I always learned they weren't allowed at the point. Pools I lifeguarded at. My kids never wore them.

But many people don't know that. I only know that because I was a lifeguard. And so I think it's really, really critical that although there is a warning on those items, that it's really incredibly clear that do not leave your child unattended. Drowning is every parent's worst nightmare. And so I think you're taking this additional step to protect kids from drowning is so critically important. So I want to thank you for that. And with that, would you like to close.

With that respectfully ask for an aye vote thank you.

We have a motion in a second. Let's call the roll item number four, AB 2146 by Assemblymember Rodriguez. The motion is do pass as amended [Roll Call] That Bill has seven votes and it is out. We will leave the roll open for absent Members. Thank you Mister Chen. You can join us if you want. Mister Low, it's your turn.

I was wondering where the mystery Chen vote came from, but you're sitting up there. We have AB 2980 from Mister Low.

Thank you very much Madam Chair and colleagues for allowing me to present to allow for local control. This is Assembly Bill 2980 and again like to thank the Committee and the hard work and accepting the amendments as noted in the analysis as well. This allows the Local Board of Supervisors and respective communities to adopt technologies and advancement of transparency and respectively ask for an aye vote and have additional witnesses in support. Go ahead.

Good afternoon. Thank you Madam Chair, Vice Chair and Committee Members, my name is Joseph Ciccolo. I'm the President and Executive Director of the California Blockchain Advocacy Coalition, or CBAC, whose membership spans the blockchain and broader web three industry. I appreciate the opportunity to provide our support and sponsorship of AB 2980 which would permit counties in California to verify or cross reference real estate records using distributed ledger technology. We also appreciate the Committee's work on this measure and collaboration on the intent.

The use of this technology has a promise of providing immense value and benefits to California residents by increasing transparency, reducing fraud and streamlining government processes, making the home buying process more efficient and most importantly, more affordable for homebuyers. California is the nation's leader when it comes to web three space and serves as a blockchain innovation hub. We are proud to call over 2000 blockchain related companies, home, of which three quarters are startups.

While the sector is rapidly expanding and providing economic value throughout the state, government applications utilizing are still underrepresented. Despite offering some of the strongest potential use cases, our organization was pleased to sponsor authorizing legislation that established the CA blockchain Working group led by Govops and experts within government, industry and academia. After a year of collaboration discussion, the working group released a report identifying the key use cases for blockchain technology, including the use of technology to build greater efficiency within the real estate transaction process.

In 2021, CBAC co sponsored legislation allowing recorders to utilize blockchain technology for the verification and issuance of vital record documents. Since passage, several counties have and continue to explore the benefits of these technologies, from urban to rural counties. In fact, Riverside County, California's fourth largest county and the 10th largest in the US, successfully piloted a blockchain solution that improve the efficiency, access and security of record keeping. Blockchain pilot projects are also gaining traction at the state level. Excuse me.

In 2022, Governor Newsom issued the blockchain Executive order to spur responsible web three innovation, grow jobs, and protect consumers. The Executive order directed government agencies to explore opportunities to deploy blockchain technologies to address public serving and emerging needs, working with the private sector, academia and community to present pilots for innovation. Following the Executive order in 2023, the California Department of Motor Vehicles, or DMV, launched a project that would put more than 14 million automobiles registered in the state on blockchain.

Early participation has shown that the issuance of vehicle titles has shifted from weeks to minutes, allowing for more secure ownership. Transfer through smart contracts enabled escrow accounts, decreasing fraudulent activity and lowering cost. As noted in one of the country's pilot projects, the best way to maintain security, efficient processing and trust is not to take the whole record and store it in the blockchain, but rather take the hash or digital fingerprint that identifies a particular record and put that on the blockchain ledger.

The intent of this Bill is to do just that. We believe this Bill provides an opportunity, excuse me, an additional opportunity for California to continue to be a leader in allowing innovative technology solutions to verify data and allow better integrations between municipal governments, ultimately benefiting consumers. Finally, let me emphasize that this Bill is permissive and in no way has a mandate. County recorders are free to decline to use blockchain technology and maintain their current systems.

Additionally, title insurance companies would also not be mandated to utilize any sort of digital ledger set up by a county. The purpose of the Bill is to reframe how municipalities can utilize emerging technology to better serve the public in an efficient and verifiable way. We thank this Committee for their time and respectfully request you join us in support of AB 2980.

Thank you. Any additional support? This Bill here today, seeing none any opposition to the Bill.

Staying on that one. Thank you, Madam Chair Members. Anthony Helton with the California Land Title Association CLTA commends the author for his work to try and improve California's real property record system. We are concerned, however, that the sponsors of AB 2980 have a fundamental misunderstanding of real property records and why the act of indexing recorded document on a blockchain doesn't provide any more meaningful consumer protection than a County record than a record recorded in County records today.

As stated in the Committee analysis, quote it is unclear how using blockchain to index records with thwart fraud. While blockchain enhances security by preventing records from being altered once filed, it cannot prevent one from filing forged or fraudulently obtained documents, end quote. The real property fraud we combat often happens long before documents are recorded in county records. A blockchain ledger will do nothing against such criminal activity.

Separately, the Bill is silent on California's extensive privacy protections for elected and appointed officials, which prohibit the public posting of personal information of legislators, county elections officials, judges and others. When responding to this concern, the sponsor is quoted in the analysis as saying, quote, the title and the deed are already public record and accessible in some way, albeit complicated and inefficient, end quote.

If the implication is that blockchain would make this information more accessible and efficient, the sponsor again demonstrates a fundamental misunderstanding of California law with respect to the privacy of public officials, including the very Members of this Committee, and makes no attempt to address this concerned. Lastly, the Bill is silent as to how county recorders, which are already struggling to make ends meet with respect to with given the current drop in transaction volumes and recorded documents, can pay for this program.

We understand it's permissive, but as the analysis points out, the counter recorders have no position on this Bill, despite sponsoring a measure recently to introduce blockchain into vital records, and we think their absence here speaks volumes. AB 2980 is a solution in search of a problem that will only provide the illusion of security and increase the inefficiency of real estate transactions today. And for that reason, we must regretfully oppose the Bill and ask for your no vote. Thank you.

Thank you. Any additional opposition in the room? Seeing none, we'll bring it back to the Committee. Seeing no questions. Okay, Mister Ward.

So, I mean, given that the recorders are not or haven't really weighed in on this, whereas they have discussed previously and for some of the examples that were given here and some of the other pilot programs that have showed potential benefit, but nobody has ultimately sort of transitioned to make this more of a permanent program like what, what's, what's. What's not working today that makes blockchain technology a superior arbiter of how to record real estate documents.

If I might, through the chair, and also welcome any additional commentary on the specifics here. But the language specifically with the Committee amends allows for local jurisdictions to opt in. So the resolution that would be required for the Board of Supervisors to come up with the mechanisms that might be appropriate. As you this Committee well knows, the 58 counties have unique circumstances in which they have their distinct nature of each local county specifically.

So the intent is to allow for local jurisdiction, should they have that capability and opportunity to utilize the technology for them to do so in an effective, efficient manner. As you know, this is also an opt in process as well.

What's the downside then, for users or for the public to be able to find information.

Yes, thank you.

Thanks. Yeah, I appreciate the question, Assembly Member. From our perspective, this Bill would allow our county recorder to duplicate their index onto a blockchain ledger title companies, given that any document in public records provides constructive notice, we and the consumers that we, our Members serve would be obligated to search that duplicative database in order to ensure that a document wasn't, you know, incorrectly copied onto the blockchain ledger that doesn't exist in regular county records. So basically, we would be duplicating our search requirements.

We also have concerns with respect to data integrity as it's copied over to the blockchain ledger. It's unclear to us how an error in the transcription of the data would affect the real property documents. How would a dispute between, if you have two documents that are ostensibly intended to be the same? However, there's an error in transcribing that data over from the county record over to the blockchain ledger. It's unclear to us how that would be sorted out. Wouldn't that be the responsibility of the recorder?

I have to imagine it would. However, I guess there's just from our perspective, it's unclear to us whether there would be some sort of impact with respect to if both databases are essentially official. How do you know which document is actually imparting notice? Okay, good.

Joe had a response as well, if the committee would like to hear.

Yeah, I guess Mister Ward's time. Yeah.

Yep. Yeah. Well, in working through this and talking to the recorders, we found out the word index has a specific connotation. So we're comfortable with the language of verify or cross reference. So it doesn't create a separate index, but rather the ability to cross reference.

Just to follow up to that. So, as I understand it, the Bill in print today does use the word index, which you're saying has a very specific connotation. And so that.

And this, I guess, is to the author, you're saying that you're open in future, if this gets out today, in future iterations, to adjust and clarify that language, to ensure that the issue that the title company has in terms of having to search two different places to ensure that there's a singular record or to do their role as a title company would do that. You're open to that?

Okay. Yes.

Yes.

All right, thank you. Thank you, Miss Wilson. Seeing no further questions or comments, I want to thank the author for taking the Committee amendments. You know, I think I started out concerned with this Bill for the reasons that were stated. The county recorders didn't appear to be here asking for it, although I guess the opt in tells me it's not likely to happen at the county level if our county official doesn't want it. Right.

The Board of Supervisors, I think, in many cases, are elected and will defer to the recorders to what they want. And so that makes me more comfortable. And then I think the amendments go even further in requiring the counties to sort of make a declaration that it is as secure and in the interest of their constituents. And I think all of that means that this won't happen unless the counties themselves decided it's in their best interest.

Which brings me a lot of comfort that we have more than the industry deciding whether this happens in each and every county. And so with that, would you like.

And also just reference also the amendment also on the sunset in 2029 as well, too. So again, the protections to allow for the state to look at the type of accountability and transparency should this Bill move forward and respectfully ask your aye vote.

We have a motion in a second. So let's call the roll.

[Roll call].

That Bill has eight votes. It is out, but we'll leave the roll up open for house Members. Thank you. Great.

Thank you. I think we only have Committee Members here right now. Assembly Member Irwin, are you okay? Great. Perfect. This is AB 2013.

Well, good afternoon, Acting Chair and Members. Thank you. I'm pleased to present AB 2013, which would require developers of AI systems and--

I think your mic, if you could move it a little closer? Sorry.

Can you hear me now? Oh, do I have to lean all the way forward? Which would require developers of AI systems and services to disclose information about their training data. Artificial intelligence has become nearly unavoidable in Californians' daily lives. New, exciting, generative AI tools are being introduced daily, and the companies who make up the cornerstone of our digital lives are adopting AI for new tools or identifying their existing tools as falling under the AI umbrella.

However, consumer confidence in AI systems has not grown at the same rapid pace as industry adoption. Many consumers have valid questions about how these AI systems and services are created and if they are truly better than what they seek to replace. To build consumer confidence, we need to start with the foundation, and for AI, that is the selection of training data.

AB 2013 provides transparency to consumers of AI systems and services by providing important information about the data that's used to train the services and systems that are being offered. This includes if certain sensitive types of data are used, like personal information or copyrighted information. It also importantly calls out if synthetic data has or is being generated and used to fill in gaps of datasets which poses special risks to an AI system.

Consumers may use this information to better evaluate if they have confidence in the AI system or service, compare competing systems or services, or put in place mitigation measures to address any shortcomings of the particular system or service. I am accepting the Committee's amendments proposed on page ten of the analysis, and now I would like to introduce my primary witness, Rob Eleveld of the Transparency Coalition.ai.

Great. Thank you. You'll have two minutes.

Thank you, Mr. Vice Chair and Committee Members. My name is Rob Eleveld. I am testifying in support of Assembly Bill 2013 as the co-founder and Chairman of Transparency Coalition, a nonprofit which advocates for transparency around the usage of training data and artificial intelligence models. Personally, I spent five years after college in the United States Navy as a submarine officer, and I believe in service. I want to thank you all for your service as elected officials.

I spent the past 25 years in technology, including being a CEO four times in early to mid-stage software companies, the most recent of which for ten years used forms of machine learning or today, predictive artificial intelligence. I've seen how these technologies can be very effective, but I've also seen how they can be abused in the ultimate pursuit of growth.

The National Institute of Standards in their language of trustworthy AI glossary has defined transparency as, quote, 'a property of a system to imply openness and accountability,' unquote. Currently, we are seeing neither openness nor accountability from the largest developers of generative AI models, and they are setting the benchmark for everyone else.

With regard to AB 2013, we are supportive of the bill providing clear guidance to the AI developers and owners because all key stakeholders deserve to understand what is being pulled into these models to drive their outputs. Some of the key dataset requirements in the bill that will shed light on the data underpinning AI models include when the data was collected, when a data set was first and last used in model training, whether a data set was purchased, licensed, or in the public domain.

Perhaps most importantly for the citizens of California, whether private information was integrated into a model. These are not heavy-handed requirements, nor will they inhibit small business innovation. Now is the time to act. We cannot repeat the missteps in action during the nascent stages of social media. The California Legislature has been a global leader in safeguarding its citizens while still empowering the tech industry to grow and thrive. I urge the Committee to pass AB 2013 and send it to the Assembly floor with a strong recommendation.

Great. Thank you. Any other witnesses in support? Seeing none, any primary witnesses in opposition? Thank you. You'll have two minutes.

Thank you, Mr. Vice Chair. Ronak Daylami with CalChamber in an opposed unless amended position on AB 2013. As we've expressed in our letter, our member companies place great value on the importance of transparency. It's how we maintain the trust of our customers and clientele and ensure the widespread uptake of these technologies.

At a high level, some of our top concerns relate to the bill's application to all AI systems and services, not just high-risk ones, and its retroactive application to existing systems and services and not just new ones. First, it would be impossible to comply with the mandates of the bill if the business has not maintained the necessary information from when a system or service was first developed years ago.

And second, just the sheer volume of systems and services implicated by the bill and the level of detail mandated in these disclosures make compliance unmanageable when applied actively. Our other major concern relates to the risk posed to trade secrets or other confidential business information. While it may not be obvious on its face, the expertise and judgment, as well as the actual selection of data and datasets chosen to train a specific AI model is itself proprietary.

We don't believe that's the intent of the author to require disclosure of any trade secrets. We really appreciate the conversations with her staff and the openness to hearing us out and considering amendments--which we're hoping to get to you very soon--and we really want to thank this Committee for the suggested amendment that focuses on a high level summary of data sets, as opposed to a description of each individual data set.

That will go a long way in helping move us closer together and resolve the concerns around trade secrets and confidential information that we are working on other amendments to fully address the concerns. And with that, we, again, are respectfully in an opposed unless amended position today, but we are looking forward to working with the author. Thank you.

Thank you. Any other witnesses in opposition?

Thank you, Mr. Chair. Dylan Hoffman, on behalf of TechNet. Align our comments with the California Chamber, thank the author for working with our coalition, and respectfully opposed unless amended. Thank you.

Thanks. Any comments from Committee?

All right. Seeing none. I'm personally interested in where the amendments are going. I don't. So, you know, we'll just have to see what happens with the Bill. But it did seem pretty broad before it was amended in terms of. I was concerned about trade secrets, but there is some, I think maybe potential.

We took the amendments and actually discussed. Them with the representatives.

Yeah. So that's what I'm saying. I think that's great. Headed right direction. So thanks for doing that, but nobody else.

May I make a comment on trade secrets?

Sure.

The company I ran processed a lot of data for machine learning scores, basically part of artificial intelligence. Most of the trade secrets are actually in the processing of the data, not the data itself. Most of these large, generative AI models are pulling in massive amounts of data, and it's much more the processing, how that data is organized, what's called out.

That's where the trade secrets are, much less what's called for in this Bill, which is just more an outline of what types of data are being pulled in.

Small clarification, do you have any concerns about if California were to implement this law and the other 49 states and hundreds of other countries don't have similar requirements?

I do not. I've been managing small businesses for most of my career. I don't think this law will inhibit any of the technology industry's ability to innovate. I do feel that transparency is going to be required for trust in artificial intelligence, number one. And we haven't seen very much of transparency in the amount of data that's being used. And also a lot of the more recent, the biggest organizations are announcing selectively already a few licensing bills.

The Financial Times just announced today they're working with they license with OpenAI. OpenAI has announced they've licensed from the Associated Press and Axel Springer Shutterstock has announced they've licensed with Meta, Google, Amazon, Apple. So they are selectively announcing things that are going into their models. We're just asking for a more consistent and broad approach to that.

Thanks. With nothing else. If you wouldn't mind opening the roll for that, that'd be great. Do we have a motion? A motion in a second. All right, perfect.

[Roll Call]

Okay. It's five to zero. We'll leave it open, and thank you very much.

Thank you.

All right, we now have AB 2863, right? I called it out anyways, whether or not you were the next one up. That's the one. That's the--I'm king for ten minutes.

That's what is says on my folder, so I think it's the right one. We're all good. Thank you, Mr. Chair and Members. Very happy to talk with you today and present AB 2863. The subscription economy is projected to become a 1.5 trillion dollar market by 2025. While automatic subscription renewals can offer convenience, in too many instances, consumers are being billed for services they no longer want. 2022 research found that on average, 42 percent of consumers forgot they were paying, still paying for subscriptions they no longer used.

In addition, the research demonstrated that consumers underestimated their monthly spending on subscriptions by 133 dollars. I suspect that we've all had stories of intense frustration when we attempt to cancel unwanted auto-renewal services. Whether it's a gym membership or streaming services or an app or online publication and computer software, the list of services goes on and the frustration grows.

If you're lucky enough to find a customer service number to call, good luck negotiating through the phone tree maze or actually connecting with a live person without the call suddenly being dropped or forcing you to start all over. We know that these systems are designed to keep consumers locked into plans.

And as the analysis correctly points out, quote, 'as automatically renewed subscriptions and services become more ubiquitous, the sheer volume of monthly and annual payments can cause customers to easily lose track of all the things they're paying for.' For the individual, many of these charges are relatively small sums, but they add up. For folks who are tight on resources, 133 dollars in additional monthly costs can be burdensome. I'm not against businesses being successful, and I frankly appreciate the convenience of auto-renew. However, if I decide to end a service, it should be as easy to cancel as it was to sign up.

And that's exactly what AB 2863 seeks to accomplish, which is it makes it easy for a consumer to unsubscribe or cancel as it was to sign up. It requires affirmative consent for an automatic renewal. It provides notice of the amount or range of the cost to the consumer, and prohibits businesses from obstructing or delaying the request to cancel.

I really appreciate the work put into the analysis from staff, and I think it summarized the issue very, very well. Here today with me to testify is California Consumer Federation Executive Director Robert Herrell.

We have a motion. Two minutes. Thank you.

Good afternoon, Mr. Vice Chair and Members. Robert Herrell. I'm the Executive Director of the Consumer Federation of California. We are the sponsors of this measure. I'd first of all like to, as the author did, call out the very thorough analysis, which I think summarizes the issue quite well. Just a couple of additional points: there is a sort of disincentive structure right now that exists in the industry to communicate clearly to consumers about what they're getting into.

And then when they get into it, what we sometimes see is what I refer to as 'the Kafkaesque maze' of getting out of it. And that is really not fair or appropriate when a consumer just wants to know, how do I get out of this if I want to end it? There's a lot of money made by that sort of, that delta, that differential, and we all have the stories, right? So I'm not going to necessarily go into detail with too many of mine.

I think everybody in this room probably has a version of that story that they could tell. Ours usually relates to, to try to save money, our nonprofit will get the deals with newspapers, digital versions of newspapers. The Chronicle comes to mind: one dollar for six months, and then, oh my God, you better not, you better put that in the calendar from six months from now because otherwise you get the 30 dollar a month or whatever it is upcharge.

This is referred to as 'click to cancel,' in many ways, the idea being it should be as easy to get out as it was to get in. That will depend, of course, on the way in which the consumer got into it. I would just also note that this is based in part on a proposed rule by the Federal Trade Commission, which has been spending a lot of time in this area the last few years.

California has every few years, kind of updated--I hate to say it, but this is sort of a high stakes game, a Whac-A-Mole that we've been playing for quite a long time in this area of the law, and I think the last adjustments were made, as the analysis points out, with a Berman Bill in 2021, and now that we have the FTC stepping up with this next thing, what this bill does is it takes that gap between where California law is and where the federal proposed rule would be.

I would just also note that the federal proposed rule makes it very clear that states, 'it's a floor, not a ceiling,' and states can go further. That's very explicit in the federal rule right here. And so happy to answer any questions, but this puts some power back in the consumer's hands so that it's easy and understandable, and with that, we'd ask for an aye vote. Thank you.

Great. Thank you. Any other witnesses in support? Any primary witnesses in opposition? Great. Thank you. You'll have two minutes.

Thank you, Mr. Chair. I'll do my best to fill those two. Robert Moutrie for the California Chamber of Commerce. Good morning, everyone, author. We are opposed respectfully to AB 2863. I want to be clear, we're not opposed to clear contracts or the ability to cancel them. Our position is based on the specific text of the bill and a close read of it. I'd like to thank staff for their thorough analysis, particularly emphasizing two points. One: we are not legislating on a blank slate here.

Consumers already have a range of protections in statute from recent legislation on this point. In line with those legislative pieces, I'll note that I've personally canceled three subscriptions in the last year, and thanks to the present legislation, have been able to do so with relative ease on phone and online. That's not to say they're not, you know, bad actors and things are perfect, but we need to recognize the floor we have. Second, I'd like to comment on the federal rulemaking mentioned both in the analysis well and by the sponsor just a moment ago.

The present text is not identical to the federal rulemaking, which is a concern to us, but also the federal rulemaking is not done. So we really see the potential for these diverging, being a very significant problem for us as we move to compliance. To the comment just made about this being a floor, not a ceiling, I think that masks the complexity of laws.

Laws are not a scale from one to 100 where you can scale up or down only on one axis, right? If language is different in both contexts and used differently, it's not as easy as saying, oh, well, this is above that number, right? It's a complex language piece. Moving to the substance of the bill, I'm just going to focus on one example provision quickly because there's a lot here and I don't have time. Proposed Section 17602-c2 dealing with phone calls is a good example of language concerns.

Under the present language, calls must be answered promptly. We don't know how to operationalize what promptly means, right? How do we come to that? How do we work with that? How do I tell my members this is fast enough or not, right? These are the kind of detailed questions that we want to work on and have spoken to the author's staff about and hope to resolve, but under present language, we have concerns about. There's another piece of language that says cancellation shall not be obstructed or delayed.

Well, I personally have benefited from when I was about to cancel a service, they offered it to me half price for meals for six months and I was very happy to take those half price meals. So in those situations, would that offer be a problem under this bill? Those are some of the concerns we have and we hope to address going forward. I also want to flag one other kind of complex piece that's not coming here is situations involving subscriptions purchased through a third party.

I've spoken to multiple members who say, okay, if this is purchased through a third party, they may have the billing information. If someone calls me to cancel, right, how do we bring those two parties together? That information together? So those are some of the practical concerns we're hoping to work through around how to make this functionally work. Again, have spoken to the author's staff and appreciate the time and hope to resolve the events, but for these reasons at this time, we are respectfully opposed to AB 2863. Thank you.

Thank you very much. Any other witnesses in opposition, name and affiliation please.

Dylan Hoffman, on behalf of TechNet, respectfully opposed. Align my comments with the Chamber. Thanks.

Steve Carlson, Association of National Advertisers. We endorse the comments that Mr. Moutrie made. Thank you.

Ryan Allain, on behalf of the California Retailers Association. Align with the Chamber in respectful opposition. Thank you.

Thank you. Any comments or questions from Committee? Ms. Wilson?

I have a question. Just a follow-up. First of all, thank you to the author for the intent of the bill. You know, I think we've all experienced these. I've had one recently where my card expired four years ago and it kept--it's still doing it and I had forgot about it. Question though is for opposition. You noted that you were opposed respectfully, not disrespectfully, as Vice Chair might want someone to do one day, but you noted that you were opposed.

You listed out things that were problematic about the bill that created it, that didn't make it quite functional in some respects, and I'm sure those are technical amends that can be figured out. However, you didn't use the word 'opposed unless amended.' So my question to you is, if those functional items were fixed, is that something that would resolve the issue or is it just something that the reason why you noted opposed versus opposed unless amended cause fundamentally you have concerns with the bill?

No, that's a very good question, and again, I always keep opposition respectful. I've yet to hear a disrespectful opposition or at least proclaimed such. Oh, was that--okay.

I welcome disrespectful opposition to some bills, but glad you don't.

Well, it was a First Amendment argument there, I'm sure, about free speech, but so I'd say, yes, it is my hope to get through all of those and come to that--and this is personal to me. I usually don't state opposed unless amended cause some Committees don't take that as a position. I just generally say oppose and then deal with the issues. So that's more of a personal touch for me.

It's not a statement of my views on the bill. But yes, we hope to get there, but I am actually convening a meeting with a number of stakeholders. There's a diverse set of interests and industries using this as noted. So I have a meeting later this week. I'm hoping to be able to get kind of a complete list and put those together. I'm hoping to get to that place of dealing with all the technicals.

The only reason I hedge is until I--and this is a problem we all have with stakeholders--until I've gotten everyone in that room and gone through all of them and had that time, and apologies for the delay in it, I just can't promise that until I know that I've gotten there, but it's my intent to, and all of the issues I'm aware of now, I expect to be able to deal with the amends. I just am not positive until I have that next meeting.

Thank you. I appreciate that. And just to note, this is one of those anecdotal--this is an issue that you anecdotally hear at the dinner table all the time. Everybody has a story about how they were stuck in a subscription or, for instance, I subscribed to all of the major newspapers, print newspapers. I've tried a few times to get out of them because I wasn't really reading it.

By the time I get off the phone with them and gone through all the steps, I end up keeping the service just to help support democracy, so to speak, but there are lots of what you would call obstructions or delays in order to get out of a service.

And people should be able to, if they would like to, opt into, I mean, if they would like to, be able to make contact with a vendor that they've chosen to be in a subscription with and cancel that subscription, and if they would like to continue conversations to see if there is an option versus, then they should be allowed to do that, have that freedom. But otherwise it should be pretty simple. If you click to get in, you should be able to click to get out.

And so I think, that definitely continue to work with the author to ensure that the technical aspects are taken care of so that it can be a functional bill and not have unintended consequences because I can see just by using the words 'obstruction' and 'delay' that there are quite a few unintended consequences associated with that. What does that really mean? Also, support the conversation around what does 'prompt' mean.

Sometimes we think we're being very specific in the bill and helpful as authors, and then it creates other issues. So I encourage you all to work with, but I can tell you that there are 120 legislators that have had this experience, and many--and we represent so many, the 40 million across California that probably had this, as long as they're adults, probably have had this experience as well. Thank you.

Thank you very much. Mr. Lowenthal.

Thank you, Mr. Chair. I wanna echo the comments made by my colleague from Suisun, and I wanna commend the author on this incredible bill that's outlying an arbitrage basically in our society right now, and so much so that there's an entire industry created right now to manage these things for us, companies like Rocket Money that are spending literally multimillion dollar advertising campaigns.

If they have marketing campaigns to that extent, you can imagine the amount of demand and money that's being made for them to be profitable on something like this, and so I think there's clearly a great need, and Rocket Money is not the only one in the game. There are so many that are in the game, and that's a growing sector. So clearly there's need for some level of relief for consumers, for them to be wanting to spend money in order to alleviate this type of thing.

And I also want to echo what my colleague from Suisun said in hoping that you can work together on this going forward because I think that there's a cost, and I say this, somebody who's run subscription-based service companies with 100 percent digital acquisition strategy online, there is a cost to not having consumer confidence in the space as well, right?

Where many people are reticent to actually joining up and signing up for services because they're so afraid of not--their lives being so complex and the technology being so complex, that there must be a gotcha in there somewhere, so I'm gonna lay off right now. And I think that there can be much greater adoption of services having that confidence in place. I would act that way too.

So not only are there 120 legislators who've been in this predicament before, there's probably 120 legislators that would sign up for a lot more things if we had a better system. So I'm hoping that this can enable the Internet economy to do even better when this is there, and I hope the author would consider me to be a co-author on this bill.

Thank you. I would love that.

And if I may, I just want to say, I really appreciate that point and I think that's true. I've had that conversation a million times, probably with my daughter, and say, 'no, we are not paying for that game because I'm going to forget that we paid for it and we're not going to cancel it.' And so I think that is, you know, it's an unintended consequence of knowing that this is kind of built into the system and people do forget, and knowing how much that adds up every month, you know, really can weigh on people's pocketbooks.

Thank you.

Thank you. Any other Committee Members? All right, seeing none, I just have a question and then a comment. One part, in order to extend a subscription, would your bill require actually the affirmative 'I want to extend?' Okay.

Yes. Yes. Thought there was more coming from--

No, no, that's all right. Most people expect more whenever. Yeah.

Yeah, yeah, right. So you would, you know, if it's a month, if it's three months, if it's a year, you would have to affirmatively confirm it, but if you confirm when you accept that contract that it will be automatically renewed, then you don't have to confirm it. So it could be, I want to automatically renew this for the rest of my life if you agree to that, then that's what would happen. So it's not limiting the automatic renewal in the way I think that was maybe misunderstood when it comes to the bill. And I don't know if my sponsor has anything on that.

Just very briefly, Mr. Vice Chair, if you look at 17602, the section in the BNP Code, part of that is existing law, and then that's being augmented as well, so that section essentially serves as a laundry list of things that you are not allowed to do as a business when you're dealing with what's referred to as automatic renewal offers or continuous service officers. And so that list is right there.

And the analysis, I think, does a very good job of talking about what's currently in law and what the additional steps being taken here are so that people don't wind up in those confusing situations. It doesn't prohibit some sort of renewal, but it lays out the pathway of clarity and clear disclosure at a bare minimum to get there.

Our main concern is that we don't want people to agree to an automatic renewal they don't know about because it's embedded in some other, you know, contract or agreement that they're doing, right, that it needs to be spelled out, that they have to know how much they're paying, that they have to know when it ends. You know, it's all clearly spelled out for them, for them to really make a decision on it.

Yeah. You know, my personal experience is most, I mean, the vast majority of them say automatically renew, and I forget, you know, every time anyways, whether I want to do it or not, but it sounds like a lot of the concerns are mostly, or, you know, there could be deeper concerns, but a lot of these can be resolved. I did have concerns about the prompt phone calls as well, and just adding some more detail into that.

You know, my staff and I were discussing this bill today, and, you know, she caught me at the wrong time because just yesterday, I had an auto-renew on something where it now was, you know, disclosed when I, you know, bought the service and auto-renew, but I didn't get that email that says that--most companies do, by the way, I mean--but I didn't get that email that says, hey, it's renewing in a couple days or whatever.

And so it was an annual subscription, too, so I might have to, you know, mow some lawns or something to pay for that, but--so, but, you know, that bothered me, right? I mean, cause I just forgot, you know, and there's everything subscription-based now, so if I have to set a reminder for everything, personally, I don't know, it's just good business practice. Most companies do it already. That's the good news. This one has a bad business practice. And so because of that, I'm going to support your bill. So--

Thank you.

But I do hope you'll work with opposition to some of that, just add some more detail on some of those things.

Absolutely. We want there to be clarity around the language, and we're happy to continue that conversation.

Great, thank you. We do have a motion. We don't have a second yet. All right, we have a second, and with that, we can take the roll.

All right. Item Number Seven: AB 2863 by Assembly Member Schiavo: the motion is 'do pass.' [Roll Call].

All right, six/zero. That one's out, and we'll leave it open. So thank you very much.

Thank you.

I think we have just one more: Mr. Lowenthal. This is AB 2481.

Thank you Mr. Chair and Members. Let me start off by accepting the amendments and thanking the Committee staff for all their hard work on this bill as well as my other bills that are before this Committee. Studies have shown severe, pervasive harm is being inflicted on California's youth through social media. Yet social media platforms continue to fail in providing a minimally sufficient response to the threats. That failure includes a generally inadequate system of responding to dangerous content.

Seriously dangerous content, including severe cyberbullying fraud, stalking, remains on platforms even after reports from at-risk targeted youth. Either the social media platform provides no response or the response comes far too late after significant harm has been done. California's educational leaders and licensed mental health professionals regularly bear direct witness to severe harms being inflicted upon youth due to content on social media, while having no means to address such content and protect the children in their care.

By creating a verified reporter process, California can enable its trusted professionals, with their expert experienced judgment, to alert social media platforms to impending severe risk facing children. AB 2481, the Youth Social Media Protection Act, will provide new protection for social media users, requiring social media platforms to provide timely responses to reported content.

It also creates a verified reported mechanism through which our trusted school leaders and licensed mental health professionals can report content that, in their professional judgment, represents a material risk to a child user's health and safety. Such reports will trigger an expedited human review by social media platform with a timely response. With me today is Marc Berkman from The Organization of Social Media Safety. I also have my Chief of Staff, Guy Strahl here for any questions.

Great, thank you. You have-- But only one testifying?

One testifying, one available for questions.

Perfect. Two minutes. Thank you.

Thank you. Good afternoon, Vice Chair and distinguished Members of the Committee. My name is Mark Berkman. I'm the CEO of The Organization for Social Media Safety, the leading consumer protection organization focused exclusively on social media, speaking in support of AB 2481, the Youth Social Media Protection Act. Current research shows definitively that social media is harming millions of California's children. About 46% of students report experiencing cyberbullying, which significantly increases the risk of suicide.

According to one study, about 43% of young adults had seen self harm content on Instagram. About 33% of those indicated they had performed the same or similar self harming behavior as a consequence. According to another study, new TikTok accounts set up by a 13 year old were recommended self harm and eating disorder content within eight minutes of scrolling the apps for you feed. Tragically, social media platforms are not taking sufficient steps to protect California's children from these harms.

That failure includes maintaining basic public reporting processes and protocols which enable the platforms to identify and respond to clearly harmful, dangerous content. The public agrees with this conclusion. About 80% of people say that social media companies are doing an only fair or poor job at addressing online harassment or bullying on their platform.

Another recent study by the ADL found that 41% of respondents who experienced a physical threat stated that the platform took no action on a threatening post, and only 14% of those who experienced a physical threat said the platform deleted the threatening content. While far more needs to be done, the Youth Social Media Protection Act seeks to increase safety for children by requiring that social media platforms review and respond in a timely manner to reported content that users deem to be a potential threat to a child.

The bill further creates a new process by which verified child welfare experts, certain California educators, and licensed mental health professionals can issue a report to a social media platform of content that poses a risk or severe risk to a child and receive an expedited review and response from the platform. This should be happening without having to resort to legislation. Unfortunately, that is not the case.

As the organization for social media safety works with schools and families across California, we continue to see educators, school counselors and child psychologists in tears over the harms they are seeing. And they are not just witnessing harm after the fact. Rather, these trusted professionals become aware of content representing an ongoing, imminent threat to the health and welfare of the children they work with.

And they must sit there helpless without any means to stop the ongoing harm, without the ability to use their professional expertise to alert the social media platform about risks threatening the children they work with. Giving these trusted professionals an opportunity to make an urgent report of harm to social media platforms is a reasonable, balanced intervention that will save children's lives. To be clear, a verification process of certain licensed California professionals will not be burdensome.

The major social media platforms covered by this bill already conduct verification processes for a variety of individuals and organizations. That includes verifying millions of nonprofits, journalists and celebrities. These California professionals are a discrete group with readily available credentials. Also importantly, this bill is entirely consistent with the Constitution and existing federal law. It does not dictate a result or certain decision from the social media platforms.

It leaves social media platforms with complete and total discretion in interpreting the risks articulated in this bill and the platform's own terms of security.

Do you mind wrapping up? Sorry about that.

Yes, the use of Social Media Protection Act simply requires a timely response from the platforms, again, for reports of imminent risk to a child. So to protect California's children. I respectfully ask for an aye vote. Thank you.

Thank you. Any other witnesses in support? Seeing none. Any primary opposition? Thank you.

Thank you Mr. Vice Chair and, apologies, but I am respectfully opposed to AB 2481. Dylan Hoffman on behalf of TechNet, again very much agree with the intent of this bill and want to note that our companies have been at the forefront of creating a lot of these new policies and features and tools to allow for greater control over the online experience.

I'll also note, it's an industry best practice to not only prohibit the many types of harmful content that are addressed in this bill, but also to allow users to report that content, as well as accounts that violate the platform's policies. However, we have several issues with the bill as it's currently in print. First, there are numerous problems that make compliance incredibly difficult, if not impossible. It's unclear how a platform will have to verify school administrators and mental health professionals for their priority reporting mechanism.

Just some back-of-the-envelope math we came up with that there are approximately hundreds of thousands of potential verified reporters in the state, and the bill is completely silent on how to validate that those people are who they say they are, that their credentials are current and have not been revoked. For example, if a school administrator moves on from that position, how would we know that? I think the bill would seem to imply that we'd have to continually verify these credentials.

Second, the bill raises numerous constitutional issues. First, the verified reporter mechanism gives a priority lane to certain government officials to request the removal of lawful content. We think that's problematic, and especially an issue that's currently in front of higher courts. The bill also sets tight timelines for platforms to substantively respond, meaning make a decision of whether content violates their terms of service and remove it or face liability.

Reports from a verified reporter must be dealt with within 72 hours or 24 hours if the report is of a severe risk. This creates a strong incentive for the removal of content in order to avoid liability. That doesn't provide platforms with enough time to go through the careful analysis to determine whether or not it actually is what is being reported.

Finally, I'll close with this. The prescriptive nature of the new reporting mechanism for certain categories of content combined with liability creates a strong incentive for platforms to deprioritize their responses to other categories of content. Since a platform can be sued by anyone, not just the victim or their parent, who reports an instance of social media related threat, platforms have very strict timelines to comply with and failure to do so will result in liability.

Platforms are strongly incentivized to prioritize resources and personnel to engineering and staffing their responses to this mechanism, resources that would otherwise be allocated to other categories of content or moderators, et cetera. This Committee passed a bill last year, AB 1394, creating a reporting mechanism for child sexual abuse material, and this bill would undercut the priority this Legislature placed on responding to that type of criminal activity. I think it is inadvisable to elevate other categories of content to that level of priority and response.

TechNet and her member companies have engaged with this Committee on similar bills aimed at providing the standard for user reported content on social media platforms. We're open to working on a bill that treats categories of content and these reporting mechanisms holistically rather than as individual or separate issues, and that takes into account the realities of moderating content at a tremendous scale. For those reasons, we are respectfully opposed. Thank you.

Thank you. So I'm late to the party. Did you both get a chance? Yes. Okay. Any other opposition in the room?

Good afternoon. Ronak Daylami with Cal Chamber. Respectfully align our comments with Technet in opposition. Thank you.

Good afternoon, Chair, Members. Naomi Padron, on behalf of the Computer and Communications Industry Association, also respectfully opposed.

Thank you. Seeing no further opposition in the room, I'll bring it back to the Committee. Any questions or comments?

Ms. Ortega?

I was just going to say thank you to the author for bringing this bill forward. Honored to support it today. And, you know, I've talked about this on this dais before. I think social media has definitely had a negative impact on our children. Unfortunately, many of us are addicted to our cell phones, and that includes our youth. And we did not put safeguards in place when cell phones came out.

And given the data we know today and the outcomes, we're not looking back and saying we wish we had put a lot of these safeguards in place, and now we're having to do things backwards. So I really appreciate your work on this bill and would love to be a co-author.

Ms. Wilson.

Thank you. You know, to the author, I know we've had lots of conversations around online things lately, especially as it relates to social media. You know, the issues that plague us in real life are exacerbated in the virtual space. And so I think that we in the Legislature have a duty to address these issues, just like we have in real life, address them virtually.

My concern with the bill is, is this the right tool to be able to address these issues? And is technology where-- And technology is pretty amazing, right, in what we can and cannot do.

But as it relates to creating the second tier of people who are typically mandated reporters or hold that space in real life, whether they hold that space online because they're easily identifiable, it is easy in real life to know whether a teacher and their authority at school is able to address an issue. Even when it comes to law enforcement, it's easy in real life to see a law enforcement and be able to address an issue. It's not the same.

It doesn't work the same in virtual, as far as our virtual identities do not cross over directly.

And given the fact that when there is an issue online, that users do have the ability to report and also block, to no longer see, and so the follow up mechanism to find out whether that material was still bad or not, you know, if that particular person doesn't want to see that anymore and wants to avoid it, whoever it is, child or otherwise, they can report, you know, and block that. And so it does protect them in that particular moment.

And I'm aware of the fact that there is, I don't know what the right thing is to call it, but for every time that happens, it creates an incident that, you know, the social media company is aware of and uses that to then police their site, like we use law enforcement to police in real life, and noting that we don't catch everything. So the goal would be in a virtual space, we are closer to 100% because there's more resources.

It's very different than in real life as far as resources. But for me, I don't know that this is the right tool in the toolbox to solve this particular issue that we, as a Legislature, need to figure out, you know, how to solve.

May I respond? Okay, first of all, thank you so much for your very thoughtful, as always, questions and comments and pointing out the concerns that you have. The first few things I want to say is that this should not be considered the only tool, but it should be one of the tools that are available to families, to mental health professionals, to school officials and so forth.

And I want to opine and ask everybody right now, do you feel that you can get an adequate response if you see something on Instagram right now that's offensive to you, just yourself? Do you have feel like you can get a person on the phone to talk to? Do you feel that you have the ability to get the service that you need?

What exact service do you--

Let me continue on, please if I may? I'm trying to respond.

I prefer if we don't interrupt each other.

On that first issue that you note, the question, to what I was saying, I wanted to make sure you clarify on that. What exactly response are you trying to get? If I see something offensive, I currently have the ability to report that offense.

How?

Every site that I-- Well, I don't know all the sites.

Via email?

Through the chair, please. Through the chair.

Through the chair. So every site that I go on, that I use, I don't use them all. But on the social media site, let's say the major three, I can immediately report the issue via their little reporting feature, not sending a separate email. And I can at that same time it asks me for all three of the major sites that I use. It asked me to block, do I want to then block that person?

So I'm asking you, you made a comment, do you feel like you can get the response you need? I'm asking you, what response are you looking for related to poor content that this solves, because the point is, I don't want to see it. So I report it as potentially breaking their rules and I block it.

Assemblymember, what I was trying to say is holistically that these social media companies have created a wall and that I was using that as an example. Whereas if you had a concern about cable, about your cell phone provider, so forth, there's actually a phone number that you can call. And I'm not suggesting that that is the way to remedy this.

I'm saying that those companies have been based on the premise that there is access, very clear access, between the end user and somebody that can actually confer with. And it has been a paradigm that tech companies have created to have a lack of contact. And as we've gotten into this crisis affecting social media of our youth, we see that even more. So in response to your question, as it relates to blocking somebody, that blocking doesn't work when somebody is being bullied in a group chat that doesn't work, how do you block that? And those are the things--

Are you talking about in a social media group chat? You can remove yourself from that group chat.

I understand if there is a youth that is being bullied, removing themselves from the group chat is not remedying the issue. When you have several kids that are in a group chat and one is being ridiculed to all of the others, simply blocking, or even if they weren't involved in it at all didn't remedy the issue.

And in school, it's the same way. If someone's being bullied in school, someone's being removed, and there's a person there who's determining that. Not the space, the physical space that you're talking about. And the more we talk, the more opposed I am to the bill now. I was more in a lean no, now I definitely am. Because just using the examples that you use when you talk about cable, cable is providing a service.

And you can argue that social media is providing a service, but it's really providing a space. And someone else, not the social media, is doing whatever it is they're doing and they're protecting-- Their job is to protect that space, so to speak. And I think that there is a mechanism to do that, meaning you as a person who's using this, determining that you want to be in this space, deciding to report and then block that person's use.

And so it's not the same example as, and I think on this particular point, what I see from your comments, you can make arguments to the rest of the Committee is that, you know, I think we just have a fundamental disagreement about social media and whose responsibility is it. I think there needs to be some level of accountability for folks using these type of services and recognizing that they are entering a public square where other people exist and other people do things that are offensive.

And sometimes what is offensive to me is not offensive to another person. But I have the ability within social media, the ones that I'm thinking of, and maybe there's a new one that I'm not, and that's potential, to curate my experience. And I think that, and I rarely find myself in being an advocate for social media.

This is weird as I'm talking, but I do think that that option to curate your experience on social media exists, and people need to do more of that versus regulating everything. We do have to protect our children as they, if we allow our children to go on social media, there should be a level of protection. But once again, I'll just note that the tool, this might not be the only tool, but tools in the toolbox. Some tools don't belong in the toolbox.

And so this is one I don't think it does. Okay, I'll end there and let you, through the chair, do whatever else you want to do.

Thank you. And if I may, Mr. Lowenthal, briefly, I had the privilege of hearing Mr. Lowenthal's daughter speak before judiciary earlier this morning and she spoke about an experience that I think is fairly typical of our teens on social media today, where they're bullied online. And she mentioned that never has she been bullied directly in her social media channel.

She mentioned that every time she's been bullied, it's been in public forum where other students and classmates can see it, and that she shows up at school the next day knowing everyone has seen the bullying. And so I guess I just wanted to highlight that point that your daughter so eloquently made today, because I don't think it is just about the user themselves seeing the bullying. I think it is about the other users that continue to see it.

Even if she were to have curated that bullying out of her feed, she would have to show up, having known everybody else was seeing the bullying. And so I think that is an important point to have made that I had the privilege to hear this morning, but I wanted to lift up again in the context of this conversation. Did you want to add anything to respond?

I'll save that for my closing remarks.

Thank you. Mr. Patterson.

Thank you. I think there are a lot of details in this bill that I do have a lot of concerns with in terms of who can make reports and those sorts of things.

But I think your intent in the bill, and please tell me if I'm wrong, is right now, you feel like if there's something online that that is a threat, you know, like a legitimate threat, that you can submit a report and maybe-- I just got a response from a report I filed who knows when yesterday. I have no idea when I reported it, and for this particular platform--

And so now I don't know how that platform, this probably wasn't a threat, but your intent is to address when there's a legitimate threat online and you want to get some kind of response to that. Correct? Is that kind of--

That's correct, Assemblymember, in an expedited manner.

Okay. And I do think we, you know, that the companies can, you know, in terms of the content that maybe I, because I agree. I mean, there's definitely content that we want to make sure, sexual exploitation of children, those kinds of things. We want to make sure those go straight to the top, and a lot of resources are put to that as well. And I think just what I'm hearing from you is that, hey, look, there's things that happen online.

What my colleague was saying from Suisun is, I think as a user, I have a lot of tools to disregard some of that content and report content, but as a person who may feel threatened or may see legitimate threats out there, we might need a quicker response to that. I agree with the intent, but I'm kind of concerned with the way that's done.

I want my, you know, counselors when they go home and the mandated reporters when they go home, you know, to be also with their families and not feel like they need to report things. You know what I mean?

I understand your concern. We're trying our best to fill a gap that exists right now, because in practice, what you have is you have situations where there is a threat, where there is a very delicate situation and a lack of response that's taking place right now from these large social media entities. I want to point out that this is something that they should really want.

They should want everyone to feel safer and that they can get a quick response so their parents and otherwise are more comfortable with their kids being aligned on those platforms. And the cost for this, despite what the opposition pointed out, is de minimis. If you look at Meta, in the last year and a half, the stock price has tripled, but there's no been a corresponding augmentation of response times associated with that.

And so I think it's far time for there to be much more responsive response times, and especially to school officials, mental health professionals that are in the trenches and need support in order to remedy any given situation. By the way, we're not trying to create new protocols. We're not saying to the social media companies that they need to respond to issues X, Y and Z. We're simply saying if it violates your existing protocols right now, does it violate your existing protocols right now?

And if somebody from a school, you know, school counselor, school psychologist, mental health professional brings that up, I think that that's something that needs to be responded to immediately. That's how I believe.

See, I agree with your intent on where you're going. I do have an internal debate with-- I don't know if I totally agree with-- I do think social media companies are interested in keeping this content off of their platform. I'm not saying they were successful in doing that, but I think there is a mutual interest in having that done. So, again, I kind of, I'm good with where your heart is, where you're trying to get with this bill.

I would love to continue the conversation should it get out a Committee. It is a tough balance, because I do want the social media companies 100% to prioritize, not only dangerous content that you've identified, but also stuff that they see that leads to dark web stuff or leads to sexual exploitation of children. I think we're in agreement on that. But I think we just need a little bit more--

Drug sales, otherwise, right? Distribution, life and death situations, right? I think we all want the same thing. And I'm not suggesting whatsoever that social media companies are bad actors. I'm suggesting that we can collaborate together with more rigid ways, requiring that dialogue to take place right away so that everybody benefits, especially them.

Thank you.

And I'll note, just for the record, that the bill is limited to companies that are over a billion dollars, which I think is an important narrowing of the bill, such that this will only apply to companies who I think can prioritize CSAM and this material. I don't think we're asking too much of companies that make more than a billion dollars. Anyone else? No more questions or comments. With that, I want to just thank the author.

This is one of several bills that you have this year really focused on our youth and their safety online. And I know it is something you are incredibly passionate about, and you are attacking it from many different angles, which I respect and appreciate because I think we all know that we have let, and you know, Ms. Ortega said this so beautifully, we have really let our kids down by not doing enough to date, and we're now playing catch up, and it's hard to do and so many kids have been harmed in the interim.

And I agree with our Vice Chair that there is a shared interest here. I think the tech companies should want people to be safe. They should be enjoyable experiences online. And I think that having responsive customer service is one way to improve people's experience in your business. And so it's disappointing that this might have to be legislative-mandated rather than just something that they've done on their own. But if that's where we're at, then I respect that. With that, would you like to close?

Thank you, Madam Chair. And thank you, Members. I appreciate everybody's comments, and particularly those that I don't see eye-to-eye with on this. I learned from those comments, and you have my promise that I will continue to work with the opposition on this. There's no desire for this to be onerous. This is desire for everybody to get on the same page so that we can step on the gas with these products and feel good about them.

At the end of the day, we must protect our children. We must give those who we entrust the day to day safety with our children with the tools to do so. And on that note, I respectfully ask for your aye vote.

Thank you. With that, do we have a motion on this bill?

Motion and a second. Let's call the roll.

Item number five, AB 2481 by Assemblymember Lowenthal. Motion is due pass as amended. [Roll call]

[Roll Call]

With that we are adjourned.