Assembly Standing Committee on Privacy and Consumer Protection

Welcome. We're going to call the Assembly Privacy and Consumer Protection hearing to order. We have four bills on the agenda today. But before we start with the Bill presentations, we're starting this new session and we have new Committee Members and one replacement Committee Member.

So I want to welcome to the Committee Assembly Members Mckinnor, Pelerin, Picci, Norris, DeMaio and Macedo. And thank Assemblymember Ellis for filling in today for Assemblymember Macedo. Welcome. And we will be voting on and adopting the Committee rules prior to starting the hearing. So, yeah, if you'll wait. We need to vote on the rules before.

We need one more Member to make quorum now that we're so large, but hopefully the Dem caucus is heading up 8. What do we half? Can I not count? We have five. We're really far away. Okay. zero. zero, we got six. Okay.

So according to the Clerk's office, we can actually start as a Subcommitee without the rules being adopted, because we are not operating as a quorum quite yet. But I'll read the remarks first and see what happens.

To effectively manage our time today, we'll be limiting testimony to two witnesses for support and two witnesses in opposition on each Bill. Each witness is allowed three minutes to present their testimony. After the support witnesses conclude their testimony, the Committee will call for additional supporters. No further testimony will be permitted at that time.

So we're asking supporters to state their name, affiliation, and position for the record. The same process will be followed for the opposition witnesses. I'd like to note that we are always and currently accepting written testimony through the position letter portal on the Committee's website.

So we invite anyone who has more to add to the record to provide it through the portal. And I will say I know that I and all of my colleagues do refer to those letters, so they. They are important to our considerations.

It is important to note that the Assembly has experienced a number of disruptions in the last few years. So in order to facilitate the hearing as much from the public, within our limits of time, we cannot permit conduct that disrupts, disturbs, or otherwise impedes the orderly conduct of these proceedings.

So we won't accept disruptive behavior or behavior that incites or threatens violence. All public comment will need to focus on the Bill and the topic being presented. Comments on other issues are out of order.

Conduct that disrupts disposal, disturbs or otherwise impedes the orderly conduct of the hearing is prohibited, which includes talking or making loud noises from the audience, uttering loud, threatening, or abusive language, and speaking longer than the time allotted. I am confident the first two will not happen. Speaking longer often happens.

I will also document on the record that the individual involved in the nature of the conduct and we may recess if necessary. And if it continues, the sergeants will escort individuals out of the building. Thank you for your cooperation and we will now begin the hearing as a Subcommitee because I do not see a quorum.

So I'm going to pass the gavel over to our Vice Chair so she can preside over my bills. And we will begin with AB412. If you don't mind, Madam Vice Chair.

Yes. While I then thank you.

Your call. You just keep you posted. Thank you.

Hi.

You can hang back there.

Not up here.

Thanks.

Any chair.

Any chair. I want to thank Assembly Member Ward for accommodating my witness who's trying to make a flight home today. Good afternoon, Madam Vice Chair and Members.

I want to start by thanking I know I'm biased, but the amazing Committee staff for all their work on my bills and on the hearing today, I'm proud to present AB412, the AI Copyright Transparency Act.

This is an incredibly simple Bill that allows copyright owners to have a right to know when their copyrighted materials are used to train generative AI. It is no secret that AI developers use copyrighted materials to train their models.

Just last week, OpenAI and Google came out publicly to say they needed to train on copyrighted materials to continue the work they're doing. And when developers use copyrighted materials that they download from the Internet to train AI, they are not obtaining consent from the rightful owner of those materials.

And the owner often has no way of knowing that their copyrighted material has been incorporated into the AI model. They aren't crediting or compensating those owners. And even when they go on to commercialize their trained models, that is still the case.

Federal law provides copyright holders with a variety of exclusive rights over their work, including the right to reproduce, distribute and display copyrighted materials. That, my friends, is a long held principle of law that someone who creates an artistic work and has a copyright has the exclusive right to use it to profit off of it.

And if somebody wants to do so, they need to license it from the copyright holder. So this Bill doesn't change copyright law. We don't have the power to do that here in the state.

What it does is it merely provides the copyright holders with notice of when their work has been put into a a data set that trains an AI model. I know last year Assemblymember Erwin, who isn't here yet but will be, I'm sure. zero, she's here. Hi.

Assembly Member Irwin passed a Bill that I don't remember the number of, but I thank you, 2013 that allowed for the disclosure and transparency of a number of things that went into the data sets. And one of those things is whether or not copyrighted material went in.

So the data sets will then have to disclose yes or no, is there copyrighted material? But this Bill really takes this to the next level and says, okay, so you have copyrighted material. We want to know, is it mine? Is mine in there?

And then when Scarlett Johansson's voice is coming out of a AI model, the people that own the copyrights to the movies and the work she was in have a right to know. Is that why it sounds like Scarlett Johansson was our copyrighted material in that work? And that is what this Bill does, as you will see.

And I. I don't know if I said that I'm accepting comedian amendments, but I am. And as you will see in the amendments, which are lengthy because I hold my bills to the same standards of everyone else, it had to be workable before it left this Committee, we made a lot of changes.

And I want to say, I know the opposition may have had time to review them all, but this will be a continued work in progress. The goal of that was really to make this workable, to make sure that the Bill was feasible.

Prior to the amendments, we didn't feel like there was a way to actually know for the data sets to be run against copyright.

And the changes we made, which allow for that hashing, which we can talk about more if there are questions, really make it possible to know whether these data sets have copyrighted material in them and allow this right to be meaningful. And with that, I will turn it over to my witnesses.

Jolie Fisher, Secretary treasurer of the Screen Actors Guild, and then JJ Simon, co founder of Transparency Coalition.

Excuse me. Before we begin, we have a quorum.

Zero, perfect. So why don't you have a roll?

Say the roll. Proceed.

They're gonna call for numbers. One second. Sorry.

Barracahan. Here. Here. Dixon. Here. Dixon. Here. Brian. Here. Brian. Here. DeMaio Ellis. Here. Ellis. Here. Irwin. Here. Irwin. Here. Lowenthal. McKinner. Ortega. Here. Ortega. Here. Patterson. Pellerin. Here. Pellerin. Here. Petrie Norris. Here. Petrie Norris. Here. Ward. Ward. Here. Wicks Wilson. Here. Wilson. Here. Okay, we have a quorum. Very good. Now we can vote to adopt the rules.

I'll move to adopt the rules. We need to adopt the rules. Yes, I spoke. Please call the roll.

We need a second on that motion. Sorry. Thank you. Thank you.

Okay. Thank you. To adopt the Committee rules. Barackahan.

Aye.

Bauer-Kahan Aye. Dixon. Aye. Dixon. Aye. Brian. Aye. Brian. Aye. De Mayo. Ellis. Ellis. Aye. Irwin. Irwin. Aye. Lowenthal. McKinner. Ortega. Ortega. Aye. Patterson. Pellerin. Pellerin. Aye. Petrie. Norris. Petrie. Norris. Aye. Ward. Ward, Aye. Wicks. Wilson. Aye. Wilson. Aye.

Let's continue then, with your. Your supporting witness.

Good afternoon, Honorable chair, Members of the Committee. This is a happy accident. We did not plan to be in matching outfits. I'm Jolie Fisher, Secretary Treasurer of SAG AFTRA and the Chair of Government affairs and Public Policy.

The union that represents 160,000 individuals who have dedicated their lives to honing their human artistry, perfecting their craft, and drawing on their unique lived experiences to entertain and inform the world. As an actor and a writer and.

And a Director and a warrior goddess activist, mother of five, butcher, Baker, candlestick maker, and a Member of a multigenerational show business family, I know firsthand that change in my line of work is the one constant in this industry. From the silver screen to home television videotapes. Remember VHS, and then Betamax and DVDs, and now to streaming.

Each new technological revolution has disrupted the old ways of doing things. The rapid growth artificial intelligence is only the latest wave of that change. And I understand that. Look, my union has no desire to halt technological progress. We do, however, want to make sure that AI improves human lives rather than harming or exploiting us.

Here's the simple truth. Everything generated by AI originates from a human creative source. AI can't do anything on its own. And no AI algorithm is able to make something out of nothing. So if intellectual property and copyrighted materials is being used to train AI models, the copyright owners need to know.

Nothing I'm saying here is particularly radical or controversial. In fact, under our current laws, individuals who distribute copyrighted material for their own financial gain can face up to five years in prison and a $250,000 fine.

Why then would we allow these big AI companies to mine others copyrighted works without at least giving copyright owners a right to know if their works were mined?

Just as a potter uses their years of training to make unique tangible objects, content creators harvest our unique lived experience to create maybe a less tangible, but undeniably unique experience for an audience. For those who make their living creating art and telling stories, our intellectual property rights are everything.

It's important to note this is only the first step in protecting humans from AI exploitation. And just as we wouldn't allow anyone to steal a potter's creation and resell those vases and bowls without paying the artisan for their hard work. We shouldn't allow these AI companies to harvest other people's intellectual property in the dark.

The AI Copyright Transparency act is sound and rational legislation. And for these reasons I urge your I vote on AB412. Thank you.

Thank you. Did you have another speaker? Please, please proceed.

Hi, my name is Jay Jessima and I'm testifying in support of AB412. I'm the CO founder of the Transparency Coalition. We're an independent nonprofit which advocates for increased transparency in generative AI. I have over 30 years of tech industry experience as a CEO, CTO and engineering leader.

I have a PhD which I completed about 30 years ago from the University of Washington and I'm also currently an affiliate Professor there. Coincidentally, my PhD was on using fingerprinting to match pieces of content against each other, which is an important foundational piece of this Bill I think we've heard a lot about.

OpenAI and Google have all said that it's game over if they don't have free right to use copyright material for free.

I think one of the things that's really important for us is to consider what we already know One is there's research done by the Allen Institute of Artificial Intelligence, a Seattle based organization that did research into some of the raw training data sets that were used by model developers.

In each of them, one is called the C4, the other one is called the Books3 database. They found ample evidence of copyrighted information in those data sets. Then the second piece of evidence you all need to consider is that copyrighted information was actually found in the outputs of generative models.

Obviously there's a well known example of Scarlett Johansson, but there's also efforts by the New York Times or the News Media alliance or many other organizations that have found that these models could regurgitate entire pieces of copyrighted content. In fact, in some cases, content that was hidden behind a paywall. And now what have these model developers done?

Rather than taking the responsible step of purging their training data of copyrighted content, they've been playing a game of whack a mole or hide the evidence. They're using techniques called adversarial artificial intelligence or prompt filtering to prevent you from seeing that they have this information in there. They're still getting benefit from this content.

It still forms the basis of their models, but they're trying to prevent anyone from obtaining evidence using the commercially available interfaces to discover that this content is in there. And lastly, I think while they're clamoring for fair use exceptions, every now and then they're engaged in a massive campaign of what I call data laundering.

They're out targeting and concluding licensing deals with a small number of large companies. And these data deals are announced with much fanfare. Not exactly a trade secret. That's not how you treat a trade secret. You don't tell everybody about it. I think the last thing I want to touch on is this technical feasibility of this fingerprinting approach.

I mean, this has been in the literature. I mean, I read papers about this in the late 80s and early 90s. So this has been around for a very long time. And one of the first things you do when you look at a large data set like this is you don't want to ingest.

There's millions of copies, sometimes of single pieces of content. And there's techniques that people use when they're processing this data to remove these duplicate copies. So it's pretty much the same technique that the industry has been using since the dawn of search to do this type of content filtering and matching.

And that's all this Bill is asking for. So, in summary, I think this Bill will do a lot of good. It'll provide us with some tools, especially for creators, to understand how these models, or if those models are using the data.

And so I'd strongly urge you all to stand up for content creators in California and anywhere and Everywhere, and pass AB412 and send it to the Assembly floor.

Thank you.

All right, thank you very much. Let's hear from the public. Any comments, your name and affiliation, speaking in support of AB412.

Hi, my name is Dina Igelsrud, Concept Art Association. We are proud co sponsors of AB412 and we encourage a yes vote today on AB412. Additionally, I am the co chair of the Legislative Action Committee for the Los Angeles County Democratic Party, which represents more than 3 million registered Democrats in LA County.

And the Los Angeles County Democratic Party also strongly supports AB412. Please vote yesterday. Thank you.

Hi, I'm Karen Gilfrey. I'm a voice actor and the Vice President of the National Association of Voice Actors. We are co sponsors of the AI Copyright Transparency act and we encourage a yes vote on AB412.

Hello, my name is Tim Friedlander. I am President and co founder of National Association of Voice Actors. I'm also here today on behalf of the Authors Guild. He was a co sponsor of this Bill and encouraged the yes vote on this Bill. Thank you so much.

Very nice.

Ortiz, how do I follow that up? Hi, I can't follow that up. Sorry. Hi, I'm Carla Ortiz. I'm an artist who has been deeply impacted by Generative AI because of its use of my copyrighted work and the work of my. The copyrighted works of my peers. And I encourage a very strong. Please yes on all of this.

Thank you.

Hello, my name is Matthew Parham. I'm a voice actor as well. Not as nice as that one, but I am a. You hear it. Director of operations for the National Association of Voice Actors and a photographer as well. And I encourage A yes on AB412. Thank you.

Sarah Flocks, California Federation of Labor Unions in strong support. Thank you.

Megan Subers on behalf of the Writers Guild of America west in support. Thank you.

Shane Gusman on behalf of the Teamsters, in support.

Good afternoon. Louie Costa with State Legislative Board of Smart Transportation Division in support.

Wednesday Ryan, a Member of the Screen Actors Guild and in full support. And it also. The Fourth Amendment of the Constitution of the United States guarantees right to property and privacy. So thank you.

Thank you. Any other comments and support? Did you want to say anything?

No, I just have a question during the appropriate time.

Okay, thank you. All right. Please come forward for opposition statement. Please.

I know it's blue day, blue day. Gotta get row a blue scarf or something.

Becca Kramer with Kaiser Advocacy on behalf of the Electronic Frontier foundation in respectful opposition. AB412 imposes an impossible new regulatory regime that would cause devastating collateral damage for research and innovation while doing little to help creators receive just compensation. We appreciate the author's work on the bill.

However, the amendments being taken in Committee today raise additional current concerns for us while not resolving our previous concerns. AB412 mandate that AI developers track and disclose all registered copyrighted works used in training is an unworkable requirement. The registration search system is very cumbersome and plenty of web content has unclear authorship. In any case, U.S.

copyright database is more like a card catalog than an actual database. Additionally, AB412 will unintentionally give Big Tech an advantage. Big Tech can afford the content licensing and legal teams to handle litigation over whether they made the reasonable efforts required by but not defined in the Bill.

Small startups cannot afford these, effectively cementing an oligopy of a few large firms. AB412 is also unnecessary. Current law is already well equipped to handle disputes over copyright infringement, with courts consistently adapting to new technologies and ensuring that copyright holders can seek appropriate remedies through established channels. Finally, the content fingerprinting amendments raise concerns for us.

YouTube's Content ID is a similar model and unfortunately it frequently blocks fair use material. Importantly, there is also a difference between a company choosing to put fingerprinting into practice as a business decision as opposed to the government putting such a requirement into law. For these reasons, we respectfully oppose AB412.

All right, thank you. Next please.

Thank you Madam Vice Chair, Members. Ronak Dalami with Cal Chamber Respectfully in opposition to AB412, we thank the author and Committee for seeking to respond to our Member concerns.

Unfortunately, based on our preliminary review, the amendments may create practical problems and unfortunately failed to address existing concerns around feasibility, interference with pending litigation, federal preemption and unraveling the Balance Policy Agreement reached at the end of last year.

Not even six months ago, AB 2013 was signed into law to require companies to provide various high level disclosures including information regarding the sources and descriptions of data sets used to train Genai as well as whether the data sets included copyright protected data.

Starting January 12026 we now face those same feasibility and over disclosure issues in AB412 yet again before even having implemented AB4 202013. Excuse me, which is frustrating to our Members who negotiated that Bill in good faith.

Moving to the feasibility issues and proposed amendments, part of the problem stems from a common misperception that all material in a trading data set will be used for training and that the data used for training will be in a format that will match a copyright holder's fingerprint of a specific copyrighted file.

In reality, there are multiple steps to processing and preparing data sets prior to training which will impact the complexity of how fingerprinting would need to be performed both to match copyright holders to content and to accurately reflect whether the content was used for AI model training.

The comparison of the proposed fingerprinting system to content ID is also inaccurate and oversimplified. Scaling such a system to the actual volume of data needed for model training and to open it to all copyright holders will necessarily create a host of practical issues.

Lastly, as acknowledged in the analysis, there are dozens of ongoing cases around AI and fair use practices, and courts are actively considering arguments that a computational analysis of publicly available works is a fair use. The Legislature should allow the legal process to play itself out.

In the meantime, it's worth noting that there are widely adopted mechanisms that help websites with copyright content to opt out of AI training. But because we feel AB412 goes too far at this time, we ask for your no vote. Thank you.

Thank you very much. Do we have people speaking in opposition? If you line up, please, and state your name and affiliation.

Yeah. Madam Chair, if I could make a couple extra comments. We're not fully opposed, we're just concerned. I want to raise a couple concerns that are. Sorry, Carl London here on behalf of the Recording Industry Association of America.

We have concerns with the amendments that were placed in Friday and we're working with the author, but there are several concerns we'd like to address.

All right, thank you very much.

Aiden Downey representing the Communication Computer Communication Industry Association in respectful opposition. Thank you.

Robert Singleton with Chamber of Progress also respectfully opposed.

Alex Suarez on behalf of the Bay Area Council, in respectful opposition.

Kelly Hitt with the Business Software alliance in opposition. Thank you.

Hi. Robert Boykin with TechNet and opposition.

Any other comments? In opposition. All right, let's bring it up to the dais. Assembly Member Wilson.

Thank you. Thank you. To the author. I understand the intent and what you're working to do, so I just have a few clarifying two clarifying questions and then comments that could be a question. So we'll let you see what that looks like.

First, can you explain the difference between the seven versus 30 days and how that would actually work? Can you give me an example of how the 7. The 7 days for if it is copyrighted, responding 7 days if it is copyrighted versus 30 if copyrighted material is not found.

I was trying to figure out how that actually would work.

Okay. Do you want to ask all your questions or no? No.

Okay.

Thank you. I'm not in charge. Like I said, that is not something I've looked at or anybody has raised. Okay, so let me take a look at that to be totally honest with you, I think our intent was to get a response in a meaningful time.

I know that we met last week with the Chamber of Technet and I'll get the communications. Thank you. Alliance together. And it was before the amendments were even done. We were working last minute. So I know that we're still in negotiations with them, but it wasn't something that was raised. But I'm happy to look at it.

We don't want the timelines to be confusing. Yeah, yeah.

Just said if they have copyrighted, it's seven days, but if not 30 days. It seems like weird to have two different. How you would actually do that. But I wondered if there was logic to it. I was like, I wanted to hear definitely.

And then one of the things that was brought up in testimony that I didn't see here and I'd heard this, but I didn't see it actually in there. And it talked about the searching and index and fingerprinting and all that kind of stuff. But on here it just says document and then it says make available upon request.

And so it seems like if it's just documented, but not. I didn't hear that. I didn't see in the law that it was searchable. And maybe that's related to the previous Bill or, you know, you're building on a foundation. And so I'm just wondering what that means. Make available to request and is it, is it searchable?

When you say document, are you talking about searchable? And one of the. And this is more of the comment. One of the things I'm concerned about when I read it is the amount of onus that is put on the copyright holder versus the developer. And so that's what I'm trying to parse through.

If you figured that out within here.

Yeah. So one of the I think appropriate, justified. Right. Criticisms of the Bill in its form before the amendments today was I think raised by Ms. Crowder. Eff, sorry, you represent a lot of people.

I didn't know who you were representing was the question around how we would make this feasible because there actually isn't a great repository of every copyrighted work. She wasn't wrong about that. And I know that was one of the criticism originally and still maybe criticism, but we have changed.

So the amendments today we worked on with, for those who were on the Committee last year, Professor Zhao from Chicago, who is an expert in AI and studies this very closely to come up with this fingerprinting notion, which for those that don't know is how Meta Google deal with csam.

So when they want to go in and find child sexual assault material, this is how they do it. They don't have a repository. They aren't sitting around with a bunch of child sexual assault material that they can look at for good reason.

They have this fingerprinting or little pieces of data that they run against whatever's in Meta, for example, to find it and remove it. Right. As we expect them to and as they want to. So this really builds on that model, which, as you heard from J, is sort of a technical approach that they use in industry now.

And so there is a shift of the onus, you're not wrong, to the copyright holder to provide that, because there was no way otherwise for these companies to really know what was copyrighted and whether it was in there.

And so they would provide the hash, the fingerprint, if you will, and then it would require the company to run it against their data set.

I will say one other significant amendment I would say as significant, if not more significant of this, which I think addresses some of the small business arguments that were made, is the Bill has been narrowed to just those that develop models.

So if you're a company that is building a model, you license OpenAI or you use llama, an open source model, and you're using a dataset, you will not have this obligation under the amendments that were put in today.

So I think that's another very significant amendment that I put in there to ensure that we're really focused on the people who are training the data sets and the people who have this ability. Because if you're training a large language model today, I think you have the technical capacity to run these hashes.

So the Recording Artists Association, we have met with them. One of their criticisms is the one you're highlighting, which is how do we do that?

So we are in continued conversations around that, making sure that we're doing this in a way that not only works for the tech companies who do this type of hashing already in different aspects, but also for the copyright holders. But I'm confident, having talked to Professor Zhao, who works with many of the copyright holders.

I know he supports a lot of the folks that you heard. He invented Nightshade, which is a tool that is used by copyright holders today to protect their works, that this is something that would be feasible for copyright holders. So I'm a little bit relying on the experts, because I'm not one.

But we're continuing to have those conversations.

Does that answer the question? It does one distal part. So the assumption is not. When we're saying documenting the search feature that you were talking about in your testimony is really when somebody makes a request, then the developer is required to search.

And at some given days, whatever you figure out as you navigate through the buildup and respond.

So the content owner would run the same algorithm on their end and they would just give them the signature. They would say, hey, I want an item. I want to see if you have an item with this fingerprint. And then the model developer would take that and try to match it against their collection of fingerprints.

So they would never be like anyone sitting there and saying search. It would just be an exchange of. It's like saying, I want something that matches this fingerprint. And then that fingerprint match has to be run by the model developer. So by the developers is the one who's doing it. Right.

Not the search and not, not upfront, not. They don't have to catalog everything. So there's objections raised around like needing to catalog everything. They don't have to do that.

So excuse me, Next time please ask the chair for permission. My apologies. Okay, thank you very much. All right, miss, you're finished. Somebody Member Ellis, did you have a.

Yes. My question is, isn't there an AI app to determine copyright infringement?

Not that I know of. I know. Do you have the secret answers on the for us? I don't know.

I'm asking.

zero, not that I know of. I don't know if anybody else at the table knows of such app.

Okay, you may respond. Thank you. I am teachable. Thank you.

So I think one of the evidence pieces I mentioned was that these models were themselves able to produce copyrighted content in their entirety. So if you said, draw me a picture of an Italian plumber with a big mustache and they would produce Mario, an image that looked just like Mario.

So the models themselves were producing this evidence and now there's some artful dodging happening where they're preventing the models from producing that output.

Did you have a follow up question?

Yeah. To my original question. I don't think you answered my question. My question is that if, and I'll back up a second in the quantum world, which will expedite the speed of, of, of AI by a billion or a trillion or quintillion.

I'm certainly concerned about privacy, but I'm, I'm concerned about this particular Bill simply because I feel like AI, if AI is that strong, then we should be able to actually determine through an AI application if there is a copyright.

Yes. Okay. So thank you, Clarification.

I think that so right now and one of actually one of the parts of the Bill is that if you have an open source model and some are open source meta's is an open source model, for example, and you provide the data set in an open way, then you actually get out of the obligations of this Bill because to your point, the Professor could see what's in that data set and find it out.

But for example, OpenAI, which licenses their tools, but it's not an open source model, I don't believe and I do turn to ro, who represents these companies to correct me, there is a way to know everything in their data set. They do, I think appropriately consider that a trade secret in a lot of cases.

And so so there isn't I think currently a way. But I do believe, I think to your point and Nightshade is the tool that I think is most similar to what you're saying. But Nightshade.

So if you put your copyrighted material online, Nightshade allows you to put something in it that affects the model and allows you to protect it that way. But it doesn't. There is not a tool or a way to know what copyrighted material is in these data sets today.

I mean, as I mentioned earlier and I think was mentioned by others, Assemblymember Irwinsville, when it goes into effect in 2026, will let everybody know for any of these models whether there is copyrighted material writ large in the data sets, but not what copyrighted material. And I don't.

But I think to your point, technology, I don't think this is going to be some onerous process that somebody is putting out. I think these very impressive tech companies will have a tech solution to this that will all be automated. That is my sense of this.

But again, we are in, you know, I make the commitment now and look forward to the continued conversations if the Bill gets out today with Chamber of TechNet, the Communications Association, whoever else wants to talk about their thoughts on technology as well as the recording artists from the other perspective.

Any other questions up here on the dice?

Assembly Member Brian thank you, Madam Chair, and thank you to the author for coming. Testimony is always interesting on this subject because we are trying to regulate an emerging and rapidly evolving industry. It's part of why the conversations from six months ago are not the conversations of today.

And the conversations of today may not be the conversations by the end of session, even depending on how quickly these conversations advance.

A lot of the testimony from the opposition at times confused me because one of the things that was said is that it's unclear what on the Internet is copyright, but also current copyright law is sufficient, and that it's cumbersome, meaning only the big companies can afford not to steal copyrighted material or prove that they haven't stolen copyrighted material.

And I think all of that, to me, actually leads towards the necessity for this Bill and the work that the author is doing. And so I will be voting for it today, strongly encouraging the author to work with the Recording Industry Association to figure out those remaining details.

And as the Member who represents sag, always good to, to see y'all up here.

So my Member, Cottage Petrinhor.

Thank you, Madam Vice Chair. And I have a couple questions about, about this as well.

If I understand, I think, I think we'd all agree with the goal of the bill, which is, I understand it is to make sure that, you know, people that are creating that that work is not sort of stolen from them effectively and then re purse repurposed in some anonymous fashion.

But I do have some sympathy for the concerns that the opposition's raised around just the reality of how we would actually implement this. So I think they raise the point that, like, there's. The standard of copyright is not sort of fixed. It's like as soon as somebody appears, something appears in the world, it could be called copyrighted.

And it's unclear how the people that have created that work can even assert that as a claim, if I understand some of the concerns that have been raised.

So I guess that feels like a really important piece for us to figure out, because it's not like if something's patented, there's some central repository that houses the fact that that's now a patented work. How. How are companies supposed to distinguish what is copyrighted, I guess is my question. And who is the owner of said copyright?

So, I mean, there is. Right. People send things to the Library of Congress to assert a copyright. That is the process. It is different than a patent. Right. It isn't. There isn't patent prosecution. And all of that happens with patents because it's slightly different. And then nobody adjudicates that until somebody has a dispute. Right? Right.

So that's sort of the difference versus a patent. The Patent Office decides in advance whose patent is. So, yeah, so you make. The point you're making is valid as it relates to the law. I guess the question to me would be, you know, this doesn't actually change copyright law.

And it doesn't, just to be clear, say they can't use that material. Nothing in the Bill says you can't train on that material. It merely says that, you know, I wrote a book or I, you know, we heard from one of the visual artists who made Marvel's, you know, scenes.

And to be clear, copyright violation can be an exact replica of the work. It can also be something you put out, a work of art that is so closely resembles the prior work that it's also a copyright violation. Right.

If I put out a book and change the names, but it's the same story that can be deemed by the courts as a copyright violation. So some of what AI is doing putting out, quote, unquote, new works could be deemed by the courts. Again, that's not the decision we're making here.

But just so we understand that it isn't even just a pure regurgitation that would be deemed by federal courts to be copyright violations. A lot of this writing, songs that sound exactly like somebody else's song could be deemed a copyright violation by a federal court. We're not determining that.

So I guess the question there, and it isn't one I've wrangled with, to be Frank with you, Assemblymember, is do we want to err on the side of more people getting this disclosure? Because again, it doesn't stop anything. It doesn't prevent the training of these models. It is just disclosure.

So if I have sent my thing off to the Library of Congress and I believe I have a copyright now, I to know if that's in this data set, I would err on the side of more disclosure because again, it doesn't stop it.

You don't have a right to a fine or payment or anything just because it's in the data set. We are not saying that. And we are not. We couldn't say that if we wanted to, but we're not. So, you know, the fair use question is being litigated in the courts and in the White House right now. Right.

I mean, the courts have decided it is not fair use. So far, it's not done. Those fights are still happening, to be clear. And we've now seen the companies go to the White House to say we want this to be considered fair use. So that fight is going to happen.

And then the question for us to ask here is, do we want to err on the side of overdisclosure or under disclosure? And I don't know. And we may have differing opinions in this room. Mine right now would be over disclosure.

I don't know that I think there's a problem in someone knowing whether their work is in there if they believe they have a copyright. But, you know, if that is something this body feels is important and wants to narrow, it's something I would be open to.

Got it. But right now. So the scope of what has to be disclosed is consistent with the scope of whatever's been, you know, filed. Someone has filed a copyright for.

I mean, the company could say, we don't think you're a valid copyright holder under the law. And then they'd have to. I mean, that is. That would be something I don't think would be impossible under this Bill because it says copyright holder. Right.

So if OpenAI says, I don't think you're a copyright holder that's not copyrighted, I guess they could. I think that that's not likely because that would be incredibly burdensome for them. I think they're likely to just run the hashes and say yes or no. I think that's the outcome.

Yeah. To be honest with you. And I think if I was a company, that's what I would do.

This is my debut on the Privacy Committee, but I think Assembly Member, and to the question of whether or not, you know, how do we balance between over disclosure and under disclosure, I do think it's important for us to recognize that asking companies to do stuff isn't free.

And so we should set requirements and demand that companies go through extra steps and jump through extra hoops because we think it creates value. Whether that's value for the creators, value for society as a whole, value for all of us.

But just saying we should not be in a place where we say there's no cost to making people jump through a lot of hoops and go through a lot of steps. Okay, my last question, because you mentioned, okay, AB 2013 from 2024, I think the opposition mentioned that as well. It seems like it's very similar territory.

In your view, What. I guess, what was the unfinished work that we did not get done last year? Like, what's the big gap that remains from last year that we need to address that motivates this Bill this year?

Yeah. So with. I mean, I was a huge supporter of that Bill, so I don't want to say that it didn't achieve the goal of the Assembly Member. I think it did. And she's nodding her head. I think the goal was different. Right. So that Bill has a long list of disclosures, amongst them being the copyright question.

But it has additional disclosures that I think are incredibly valuable as well that are unrelated to this Bill. Wholly. So the Question for me, and this is something the one thing, I mean, I've said this before, and I'll say it again, is I look at what I We should be regulating in the copyright.

Sorry, in the AI space. It is the principles that we as society have held for a long time that now are different because we have a different way of. And you'll see my notification Bill is the same thing. It's something we've all agreed is wrong and how do we vindicate those rights in the new world of AI?

And this feels like that to me. And so although that was an absolutely great first step and again, I supported on copyright, that was one small piece of what assemblymember Erwin was trying to achieve.

And this sort of goes further into the copyright space to say where this right, that again, we have long believed is one people have a right to. How do we meet that out in a world of these data sets that are black holes and where we see these works coming out that look a lot like copyrighted material?

Well, did the model make that up on their own? Because it may be that we do. The other interesting thing about this, it may be that we think it is a copyright violation. I hear a song that sounds just like somebody else's song and I'm convinced they trained on Taylor Swift songs and it's a copyright violation.

It may be that data set has no Taylor Swift in it and the algorithm created a song that sounds really similar. So it could go both ways. Right. We assume that it's one thing, but it may be another. And so the ability for that copyright holder to go and find out, I think is critically important.

And I agree with you. I don't think that we should be putting burdens that aren't meaningful.

And yet I think I believe so deeply in copyright as this right that has created the works that each of us has been better for, the books that we've read in school and that have broadened our minds and our viewpoints, the movies we see that allow us to get into somebody else's shoes and think about the world in a different way.

And I believe if we create a world where that is no longer valuable, we will not have that anymore. And so I do think this is really. I personally think this is really important. And the reason I think over disclosure isn't necessarily more burdensome is because I think this is all going to be automated.

And so I don't think that if you allow slightly more people to try to express this right, it's going to be more burdensome. Just because I think they will have to create the system in the first place. And then the number of people seeking this answer won't change the economics for the companies.

Got it? Yes.

And I think I would say, and I, you know, I'm going to be supporting the Bill in Committee, as I said at the beginning, I think we'd all agree with, as you said, that, you know, the artistic geniuses who have created, whether it's a song or a novel or a piece of art that have inspired us and touched our hearts.

Certainly we've got to make sure that in this new frontier that the principles we've established in our current world are carried forward.

And I do think, because we're grappling, as Assemblymember Brian said, with technology that's evolving so, so rapidly, figuring out how do we right size regulations to make them implementable and impactful feels like kind of at the heart of what the journey of this Bill will be as it moves through the Legislature. Absolutely. Thank you. Thank you.

Thank you. So many. Member Patterson.

Thanks. Appreciate it. Three kids sick from school today, so haven't slept for two days, so I'm a little grumpy today. Apologize in advance.

Mine are all in school today after weeks of sickness. So you'll get there okay.

Yeah, someday. Someday it'll let you know. They say I'm gonna miss it when they get older, so probably not this part. You know, I think, you know, last year, you know, there's been legislation served on this Committee for a while.

You know, obviously taking data from, you know, like, like an actor or something and then replicating that and then who gets compensated for if that person isn't actually acting. Right. It's the AI is recreating the voice. And those were interesting discussions in prior years, I think, and I'm always open to that.

But I, I actually, just to be like, completely honest with you, just between us. Yes. Yeah. Yeah. I think I actually have a concern with, like, the underlying premise of this Bill. And the reason is because I'm not sure I'm. I'm really there. And it's not this Bill.

It's the idea that we are going to constrain AI, and this is a very specific instance, but we're going to constrain AI to a point like whose data set should we take out or should we disclose or. I mean, these are very broad questions and they're fair points from the proposals opponents, by the way.

I mean, I'm not dismissing any of those, but this is a very broad conversation and I Don't think we can really look at it, at this. This detail because, you know, we talked about the potter person making pottery.

I mean, that person likely, like all artists inspired by somebody use somebody's, you know, somebody who's painting something, you know, may have been inspired by Leonardo DA Vinci and, you know, used inspiration from that painting and actually concepts from that to make their own painting. In fact, you're taught to do that in school.

And so when I look at it also from that lens, I think that that wouldn't require disclosure on the new artists who do that. And so I think we actually rely right now on AI taking all the data they can get to develop the answers or the things that we're looking for. So I'm not really sure.

It's hard for me to get around the idea of taking one particular data set and limiting it, because I'm assuming that a lot of the data that AI takes in is copyrighted somewhere. And in fact, the courts are figuring that out right now.

And I'm kind of desire for those to sort themselves out somehow rather than us going on our own. But I'm happy to obviously listen to any response. I'm sure you have a response on that. But. But I think this is a very small portion of maybe a lot of data that goes into these systems.

And we could have 100 bills just on that particular issue with some other group next hearing.

So want to reiterate one thing, then I'll respond. And if you don't mind Ms. Fisher weighing in, Madam Bistro, that would be great. But I think it's important to note this doesn't ban the use of any material, period.

If the courts find its fair use, then this Bill wouldn't give them a right to any compensation for that training. And there's also two things happening. So one is, you know, the.

The AI companies are claiming that they're training on, say, a movie to teach how language happens, and then they're just using that to, you know, train the model. How language, just like watching a movie, to your point. And the claim is being argued whether that's fair use, again for the courts to decide.

And I think that is an interesting legal question and one we don't need to wrangle with here because we don't get to decide. But then there's the second question. So that's the inputs and the training, then there's the outputs. Right?

So in your example, where it's just training and it's learning how language works and it's outputting A conversation on whatever. Something random, how to cook tea, you know, pasta.

Then, like, that is a very different thing than if I go on in the example that was given by my witness and I say, draw me a picture, you know, that looks just like a Leonardo. That's not copyrighted anymore because. But, you know, or Mario, create. Create a song that sounds like Taylor Swift.

Or create a song, you know, describe Taylor Swift and don't even say Taylor Swift, and it outputs something like that. That, by all means, we would agree, is a copyright violation, and the courts would, too. Again, not a fair use question.

That is actually, you cannot put out a song, you know, or book that is exactly like someone else's. That's not allowed. So, you know, I mean, I just think there are different copyright questions to be posed before a court and not for us to decide.

But the information about whether your copyrighted material went into the training that gave that output, I think is a valuable one. But again, differing minds. You know, I'm not. I'm not here to tell you that you don't have a right to your opinion on this by any means, as you know, Mr. Patterson. But I did want.

I think Ms. Fisher wanted to weigh in, if that's okay. May I, please.

Well, I'm sure that you all heard that we had a once in a lifetime, once in a generation anyway, strike, hopefully once in a generation that we came through with flying colors, were able to achieve very specific nuanced language and generative AI in the 11th hour.

It was literally 1107 when we were able to get that language in the contract, always knowing that we needed legislation to enforce some of the things that we talked about, and our studio partners agreed to.

My sister Carrie Fisher was able to star posthumously as Princess and then General Leia, because there was transparency, because there was consent, and because there was compensation. It was not true of Peter Cushing, who was in the very same movie in Star Wars.

So that's just one film that we were talking about that we were able to train models. They were able to use a synthetic performer, which is, like, terrible to even have that come out of my mouth, that we have that. But what we're specifically talking about is training models with multiple pieces of art, with multiple people.

We could all be put into one of the blenders, and they could spit out a character that is a little bit of all of us. What we are trying to know is who those things, who those. What the training system is taking in, and that I think that the word transparency has become a dirty Word.

And I think that that's really what we're talking about here. We're talking about what is going into these training models and being able to say, hey, there's a little bit of me, there's a little bit of rock and roll and a little bit of this. You know, I think transparency is not a dirty word in this case.

Yeah. And in the case you, you started with, you know, replicating the actual person, I think is, you know, those conversations I'm always open to have. And, you know, just one time I was messing around with my buddy Grock and I asked Grok, I don't know if Grok is male or female, but he's a friend of mine.

But I said, hey, can you make me an image of Governor Newsom doing something that I was going to post? I thought it was funny.

You created a deepfake Assembly Member.

Well, I know. So it's about. Well, there are all those laws. But, you know, but the thing is, presumably the image of the person you're asking to make of is coming from some kind of material that is probably copyrighted by a photographer, because not like rock is out there taking pictures of people or whatnot.

And so I guess theoretically, you know, should there have been some disclosure like this picture was developed from, you know, picture on the, you know, the SAC B picture, photographer, or this. I don't, I mean, I don't know. How do you use. How do you even use AI, if that's.

Or especially on the image front under the, under this Bill, maybe it provides some disclosure or something. But I just that presumably that information came from copyrighted information. And are we. And while it doesn't prevent this, this Bill doesn't prevent it as amended. I'm not sure I see the utility in knowing that.

It used a whole inspiration of data to come up with that.

And just 1.0 of clarification on that, which is you as the user would not see any difference after this Bill and before this Bill. The difference is that, you know, someone who is a registered copyright holder could, through this process, seek information around what was in the model. So it doesn't.

And to be fair to Grok or, you know, my boy, it may be, you know, Governor Newsom puts pictures out all the time that are not copyrighted. You know, so that's another example of where, you know, he's putting pictures himself on Twitter that are fair, not. Not fair use in the copyright sense, in the larger sense.

It was a very handsome picture, I will say.

So he probably put it, I Don't know. He probably voluntarily put that on the Internet.

Thank you. But he is a public official, so that it is.

And again. Yeah, yeah.

Yes, sir. Yes.

Just one comment. I think, you know, we tend to anthropomorphize models we think they think that they create. But really the way to think about these models is like, it's like if you're familiar with a jpeg, which is what a photograph is, it's like a fuzzy JPEG of all the data that's in there.

So it's only producing things that are a variant or an exact replica of what's in there. So it's not. Industry likes to use words like reasoning or things of that sort. It's not the same way that we would use reasoning in a human context or creativity in a human context.

The models, the way they work mathematically, it's always about looking up patterns and producing things that are based on the patterns that are already ingested. If it's not in there, it's not coming out. Okay, thank you, Sami.

Member or did you have a question?

Thank you Madam Chair. Thank you Madam Chair for introducing this Bill.

I want to lots and sad and I maybe as a point of hopefully summarizing some of it, but also simplifying some of it, I wanted to start off and say that I think we agree very strongly on the intent and the purpose of this, which is that if somebody is working to be able to create some kind of a content and that content is then advertently used for some other purpose which is might even lead to some profit making activity that runs completely afoul and we deserve to make sure that that's corrected.

And I think that sort of the principle, one of the driving principles that you're getting in here is to correct this oversight in this new era of gen.

So 100% in alignment there and then thinking how does this play out in real world is where we get like, you know, down all these rabbit holes of how is this going to actually work? Because we want it to work. And if the law already prevents copyrighted material. Right.

From being used and I think that this is going to be further enhanced. You had mentioned under AB 2013 as well. How would just then the improvements that you're trying to make here through disclosure and documentation afford benefit to the content? You the creator. Right.

How are we actually helping them make sure that they are paid for the work that they are doing and any sort of, you know, sort of secondary use of that work? Right.

So I think again, I think the first step is a really important one, and maybe one we consider. If, you know, the large language model company under AB 2013, I'm gonna get the number wrong too. You know, says no, then I don't. You know, they clearly don't have any copyrighted material.

But my guess is every single one is gonna say, yes, I might be wrong, but I think that our sense is, and the companies are coming out and saying this, that they are using it. So then the question becomes, how do you vindicate it? Right now, Ms.

Ortiz, who you heard from, who is an artist who, as I said, created the Marvel movies and other magical things that you've seen, is part of a union and they were able to sue. Right, for copyright violation because she was seeing her work spat out of these models.

Now the union has the resources to do that, and they were able to vindicate her rights through a very lengthy and expensive lawsuit. Imagine if you, Carla Ortiz, and you're not part of a union and you're an artist who is able to get by in California. But that's just about it.

You can't sue to get discovery to figure out if they're violating your copyright.

And so this is a very simple way to say that someone like that could go online, go to a website, put in their hash, and figure out if that copyright has been violated such that then perhaps they'll go forward in vindicating whatever rights they believe they have. Does that make sense?

It does. Would you characterize this Bill then, as a way to be able to maybe put an additional check and balance on the responsibility of the Genai companies to basically check for whether or not they're using copyrighted material?

Yes. I mean, I. But they won't. I don't know. I mean, if still forthcoming, they won't. You know, the reason we added this hashing in is because it is very hard to know up front if they're using copyrighted material, unless they're intentionally pulling copyrighted material. Right.

If you're going on the New York Times and pulling everything you have, okay, you know you're getting copyrighted material. Right. I mean, there are ways to know it, but sometimes when you scrape the whole web, you may not know it.

So I don't know that this Bill is going to solve for the LLMs knowing up front that they are violating. But I do think that if, you know, some output were to show a violation, and I could go on and say, well, did you train on my work?

Well, if I get a no, I have no right, because clearly it's not a copyright violation. So I can walk away happily and know, wow, that math worked in a way that's really fascinating. If I get a yes, then the question is, did it violate copyright law and do I have rights to vindicate?

And if so, maybe there's a way to deal with that without costly litigation. Also because again, in the world of copyright, you don't have to never use somebody's copyright. I mean, the example that was given by Ms. Fisher about her sister is you can use people's image as long as they are paid and compensated and licensed.

And so maybe this is a way to say, okay, now I know you use my work, I'd like you to pay me for that. Right. Maybe it's a small amount, I don't know, that's up to the license holder.

Right. And I think things to think through as you're continuing to work on this is the inadvertent use or maybe even the sort of the secondary use if you have a copy of a copy that no longer has that fingerprint attached and you are inadvertently starting to upload this into Gen A models.

And now as somebody and trapped over here through either this Bill or through pre or through existing law, through no fault of their own, but the content creator wants their due compensation as well. And how do we reconcile all of that? Right. Yeah.

And that is one of the reasons that I made the change to just focus on the large language model developers, which I appreciate it takes you back to the original source, not put out the resources.

If they're that big that they can. I get that. And I think that's a very good amendment.

I would echo some concerns that I know weren't able to be made in testimony today from the recording industry and also on behalf of many of their artists about whether or not some of the amendments might be shifting some of the burden. Right.

And so that's something I think to be able to continue to work on, you know, in subsequent steps as this moves forward. But you know, I like where we're where the foundation of this certainly.

And I like, I think where things are headed because I know you're a very diligent author, a great chair, and we'll continue to work on all these issues. It's complicated when you're working in this new subject area to try to solve something that in principle is so simple.

But of course in the world we live in right now has all these other technical things, considerations that we have to work on too.

And I'm always optimistic that as the process goes on that we will hopefully find a technically feasible solution. We were hoping this was it, but obviously open to conversations both with the copyright holders and with industry to try to figure out a way that makes this not, as mentioned by impeachment.

Do we have. Did you have any more. Are you finished? Any more questions up here? Yes, Ms. Pellerin, one more.

Okay. So as the bill's written, it applies only to copyrighted materials which are registered with the US Copyright Office. So do you see any registration working retroactively? So I'm thinking of a company doing their best to exclude registered copyrighted material and then being penalized when that somebody copyrights after they've already started doing their training models.

Does that make sense?

So a work that is put out into the public that is not yet a registered copyright. So now I've trained on it. That one didn't come up either. Thank my staff for this one.

Yeah. No, I love.

This is why I love the Committee process. You know me. So I never mind these things coming up. If they didn't come up, up for. So that. I mean, I wonder how often that happens. Usually people do register their copyright before they put it into the public sphere in order to protect it.

So I don't know how many works that would be, but I guess we could look into that because I actually was asked a question earlier about the fact that we don't have a right to delete here, which I think is something we didn't do, which would be incredibly burdensome because you can't actually, once you've trained the model, delete data without deleting the entire model, which then would mean retraining the whole model with the data set, which for talk about unduly burdensome.

The cost of training these models is. I mean, there's a lot of energy and water and things that go into training them. So, you know, we don't want it to be burdensome. So. Yeah, no, I will. Yeah. Thank you, Ms. Cutler. We'll look into that.

Any other questions on the dais? Am I okay looking at everyone?

I mean, we've talked it.

No. Okay. Do we have a motion?

zero.

zero, we do. Your closing statement respectfully ask your.

I vote. Thank you. Madam Vice Chair. We have a motion. Second.

Second. By Ortega and Lowenthal, please. Thank you. All right, let's call the vote.

Okay. Item number two, AB412 by Assemblymember Bauer Cahan. The motion is due. Pass as amended, to the Judiciary byrocahan.

Barra. Cahan. Aye. Dixon.

Aye.

No.

[Roll call]

[Roll call]

The vote was nine. The vote was 92. And we will keep the roll call open. Thank you. All right, Mr. Ward, looks like you're next coming Member Ward all right, it's all yours.

Want to get AB 446 when you're ready.

Thank you, Madam Chair and Members, I first want to thank the Committee for all their work on this Bill, worked really closely with my staff and myself to be able to understand and be and be able to provide some of the improvements that were before the Committee here today. So I will be accepting the Committee amendments.

I do want to note that we are continuing to work on a few concerns that are raised with the Committee. One, we'll make sure that we're going to ensure that nothing in the Bill would preclude insurers from being held liable under Proposition 103 or any other provisions of the insurance code.

And in addition, we're going to continue to work with the judiciary where it will be heard next should it pass this Committee on additional amendments to help on some definitional amendments about replacing terms prevailing party to prevailing plaintiff in order to correct some of the drafting errors and make sure attorney's feet are properly awarded.

And I'm sure there will be other things that we work on because this is a very, very interesting, exciting and new phenomena that we are experiencing as consumers. And we're trying to correct a problem here through Assembly Bill 446.

What we have seen is a growing body of evidence, and a lot of this is actually from investigative journalism that is starting to notice that different consumers are paying different prices based on your own personal characteristics and who you are.

How are they identifying that is really through your own electronic information, whether it's on your phone or on your home computer, on an IP address or some other sort of electronic information that is deciding that you fit into some kind of a characteristic about who you are.

And maybe you are a male, you are a female, you are from a different community, you live in a different part of the country, and therefore you should probably be charged a different price for a given product of what you want to purchase.

And you don't even know that this is going on sometimes until say, somebody's out there, maybe asking somebody else of a different demographic and a different location. Hey, can you go online and be able to identify whether or not you what price you see for the very same good or the very same service.

And so what we have is a new experience in consumer affairs called surveillance pricing, that is using your own personally identifiable information through your own electronic communications to decide what price you should pay. And this runs completely afoul to how we have always experienced opportunities to be consumers.

You know, might even go back to the days of bartering where different people are paying different prices for the same things. But we have always believed that there should be one price for the one good for all people.

And of course, we're going to get into some other permutations of this as we talk about discounts in other areas as well, about where the General public may be afforded some ability to have savings or differential pricing. And there's a distinction too. You're going to hear a lot about, you know, supply and demand and dynamic pricing.

And none of that, I think, is affected through this Bill. But I'll just, you know, in the summary of a lot of these studies, I'll really highlight one that Consumer Watchdog published last December, which noted that in through expose in December 2024, that businesses have really been aggressively using. Now they're becoming aware of this.

In fact, they've been encouraged, and I'll have evidence for that as well, to use surveillance pricing so that you can have different pricing for different people based upon some of their individualized data.

If you go to my home county in San Diego in 2022, the company target actually was already found guilty by our District Attorney and had to pay a fine of $5 million because they were trying to charge individuals a different price online, getting them into the store, and because of their geolocation data on their phone, realized that they were inside the store and automatically changed the price.

Now that fortunately was already solved under existing law because they were entrapped. They were. It was essentially a violation of consumer fraud law and they were brought into the store, but it exposed. What's going on here is that we are starting to see this practice of connecting our own data to make differential decisions.

That's compounded now with again, the era that we're in. Companies existing ability to be able to collect and analyze massive amounts of customer data, the installation of facial recognition as you walk into a store, electronic shelving labels as you go to approach the product that you want, and that could actually change in real time.

And trying to make sure that you are linking that to a customer's individual profile, we believe, I maintain that this is predatory, this is discriminatory, and it violates public trust at a time when consumers are already stretched thin and they don't deserve to be unwillingly exploited.

Additionally, evidence of surveillance pricing is also suggesting that it disproportionately affects lower Income, individuals and those with fewer shopping alternatives. So ensuring fairness in pricing is not just about economic justice. It's about preventing a new form of digital exploitation.

As you all know, because we work hard in this Committee, California already has some of the strongest privacy laws in the country that give consumers control over the data. But this emerging practice appears to exist through loopholes that allow companies to charge different prices based on that data.

So AB446 will prohibit the practice of surveillance pricing by making it unlawful for businesses to use personal data when charging different prices for the same product or service, whether it's online or doing an in store checkout. We believe that the right to fair pricing should not be a privilege for few, but a fundamental protection for all.

And with me to speak in support of this Bill are two of our co sponsors. We have San Diego so Consumer Watchdog President Jamie Court and Kristen Heidelbach, the Legislative Director for United Food and Commercial Workers Western States Council.

Thank you. I don't think I could have said it better. So I'm not going to say much. I'm going to say this. Companies are using our data against us. Orbitz charged Mac users more because they use Macs and they knew they would pay more.

We found Lyft, two people going the same place, same route, Same time, charge one $5 more. Happens all the time. The Target example is an interesting example.

It's an interesting example because there is a law that says if you're, if you have a store, something posted, a price posted on a shelf, everybody in that store has to be charged that same price. So the reason that they got this target paid $5 million was because they were charging people something different on their phone.

They went into the store and picked it up and they violated that law. We need a law that says whether it's in a store or not, if you have one product, you have one price and they can't exploit the data they know about you to get a different price.

The Federal Trade Commission revealed even more than we know now. But suddenly something happened and the Federal Trade Commission stopped investigating. I don't know what happened. Something happened. The new President took, took office and the Federal Trade Commission pulled the investigation. Well, you have an opportunity to do something that they can't do now.

You can say no to surveillance pricing. We can do this in the State of California. And that's why I urge you to support AB 446. Thanks.

Thank you. When you're ready.

Thank you. Good afternoon, chair and Committee Members Kristin Heidelbach with UFCW Western States Council proud co sponsor of AB 446 thank Assembly member Ward for carrying this important measure. UFCW represents workers across California, Arizona and Nevada in grocery and drug retail stores.

Working in our local grocery stores and community pharmacies, our Members see how technology is being used to maximize profits at the expense of consumers and workers. If a customer thinks the price is different than what they were rung up for, as Jamie mentioned, our Members are the people that they blame it on.

They're the ones they complain to. Today, retailers are putting more and more effort into knowing everything that they can about their buyers, their preferences, how they use different products or services, when they shop, when they get paid, and what they're willing to spend more on. The list goes on.

Retailers already have the tools they need to put surveillance pricing into effect. They have the data. They have the algorithm and the infrastructure. Grocery companies are investing more money into AI technology in their stores and looking to install digital price tags on stores.

Which means then it wouldn't just be a practice online, but it could potentially happen inside of stores. Paired with their existing ability to collect massive amounts of customer data, these technologies give companies the ability to change prices in real time according to a consumer's individualized data profile.

If a company knows you get paid on a Friday and you go grocery shopping once you're off work, they might raise the price of bread for you by 50 cents. Customers deserve to be charged the same price for the same product.

As surveillance pricing continues to mature and deployment spreads, it is important for researchers, enforcers and policymakers to keep pace with new technologies. It's critical to lay out clear principles that guide future developments in enforcement in this space.

That's why we're supporting AB446 to prohibit businesses from using the personal information of a consumer, to adjust the price of goods based off their individualized data and ensure consumers have the right to a standard pricing option.

Without legislation to protect workers and consumers from companies that hold all the power and resources to set prices of everyday items, prices will continue to rise and push Low income families further and further behind. Thank you, we urge your aye.

I just want to say one thing about the amendment the Committee proposed on exempting insurance companies totally and I understand we're going to work it out. The reason that wouldn't work is because two insurance companies have already been held accountable for surveillance pricing based on violation Prop 103.

You can only price based on driving safety record miles driven years of driving experience and two insurance companies were caught all state and farmers and enforcement actions actually looking at how willing people are to pay more prices, higher prices. And all state customers who are willing to pay higher prices based on their data were charged higher prices.

They were caught, they were penalized. And we don't want anything in this law to conflict with existing law. So we appreciate the amendment, you know, changing to accommodate the fact that there is a law preventing the surveillance pricing. We don't want something that would exempt the companies from surveillance pricing without recognizing that existing law. Thank you.

Thank you. And I always appreciate the author and his willingness to work with everybody. Yes. Yes. Anybody else here in support, Name, organization and position, please.

Good afternoon, Madam Chair Members. Robert Harrell, Executive Director of the Consumer Federation of California, in strong support of this measure. Thank you.

Marika Yoshihara, asked to express support on. Behalf of Western Center Law and Poverty. American Economic Liberties Project, Oakland Privacy and UDW AFSCME Local 3930. Thank you.

Thank you.

Kimberly Rosenberger with SEIU and strong support.

Hi. Samantha Gordon with Tech Equity and support. Thank you.

Madam Chair Members. Sarah Flocks, California Federation of Labor Unions, in strong support. Thank you.

Thank you.

Becca Cramer-Mowder on behalf of Kaiser Advocacy here in support for Privacy Rights Clearinghouse and also California Low Income Consumers Coalition.

Good afternoon, Chair and Members Marissa Hagerman with Tratten Price Consulting and support on behalf of Economic Security California Action and Consumer Attorneys of California. Thank you.

Thank you. Navneet Puryear on behalf of the California School Employees Association. In support. Thank you.

Nicole Rocha with Tech Oversight California. In support.

Mitch Steiger with CFT, a union of educators and classified professionals. Also in support.

Madam Chair Members, Louis Costa. With the State Legislative Board of Smart. Transportation Division in support.

Hi, Madam Chair Members, Jp Hanna with The California Nurses Association standing in support. Thank you.

Thank you. I think this is our opposition. Yes. Okay. No, I just didn't. I was like, I don't think you're coming up to meet you and support. Although Mr. Ward would welcome that, I'm sure. So three minutes each. You may divide it as you see fit.

Thank you, Madam Chair and Members. Robert Moutrie for the California Chamber of Commerce. And we are in respectful opposition to AB 446 because we see this fundamentally conflicting with portions of the CCPA that are in effect and making it harder for us to offer customer friendly programs that we presently do offer.

Before I get into substance, I want to thank Committee staff for the really detailed analysis going through this tough topic, really distinguishing the history of pricing and also different types of pricing. Right. Dynamic pricing versus surveillance and other pieces. That's not an easy topic. Also appreciate the on going talks with the author staff.

As recently as just a moment ago, my testimony today will be focused on the proposed Committee amendment text. I don't want to waste everyone's time with the prior text because they're being taken. Obviously we view the prior Bill, which was a straight kind of ban on all discount and loyalty programs, as more problematic than present.

But we do have concerns with the direction gone and I'll get to them here and I want to be first clear and deal with the anecdotes flagged. We certainly do not support some kind of discriminatory pricing.

In our letter we flagged that, I believe most of the anecdotes cited from the Consumer Reports publication we do not believe are present or we believe are outdated or incorrect. We will continue to work to show that evidence.

I've spoken to the staff on some and we'll continue to deal with that because we do not support those things but also are not aware of them happening. What we do think on this amended text, I'm going to flag one CCPA conflict. We think not a loophole, but the conflict.

Cal Civil code 1798.125 B3 provides a business may enter into a financial incentive program only if the consumer gives the business prior opt in consent that clearly describes the material terms of the program which may be revoked by the consumer at any time.

So when we say that is the law that presently governs these loyalty programs, we do not see that law as incorrect and we see this as in conflict with that changing those standards without removing that text from law, which for our Members creates kind of a conflict confusion of is the CCPA governing how this privacy is handled for our programs or is it, you know, in this other statute here?

I apologize for not flagging that sooner. It just wasn't in the Bill till Friday night effectively also flagged. This Bill treats CCPA data, aggregate data, I should say under the CCPA, as if it were personal information. That's a change which without spending more time on it, we view as problematic.

And third, I want to flag I appreciate that insurance was mentioned, but we believe there are other industries this poses clear problems. For example, I've spoken to lenders in banking on the idea of certain transactions are inherently based on personal data. I realize I'm at time, so I'll be very brief there.

So those are kind of a brief list of concerns for this very complicated Bill. Appreciate the time. We'll continue to work with the author.

Great and he saved you 20 extra seconds. Go ahead.

Oh, thank you. Thank you, Madam Chair. Thank you. Assembly member Ward. Daniel Conway, on behalf of the California Grocers Association in opposition. First of all, I want to acknowledge the amendments from Committee staff. We feel like it was a positive step and again, appreciate your office taking those.

I just kind of want to reiterate some of the points that were brought up around some of the concerns we have with the Bill. Obviously, for the grocery industry. We're in the business of feeding people and we use our loyalty and rewards programs to give our customers what they want frankly. You know, we're a family that we buy a lot of yogurt, we buy a lot of produce, we buy a lot of things like that, and we get coupons that reflect what we shop for.

Even with the amendments that we've seen today, we still have concerns about our ability to personalize those type of coupons and discounts for our customers. Namely the language you have around offered to all consumers on equal terms. We're still struggling to figure out how you can do personalized couponing with language like that in the Bill.

Beyond that, you know, we do service customers who utilize WIC and SNAP and programs like that. And a lot of our grocers provide incentives for those types of customers too, like free delivery or like half off rewards programs, things like that. And so we, we are concerned that those customers would probably unintentionally be harmed by this Bill.

My colleague here raised some of the concerns we have about redundancy with the CCPA. We think that that gives customers the option to consent to being part of these programs. And we think that that language has done well to date.

And we think that this is kind of going to add just a redundant process that's going to frankly, make more people probably opt out of a rewards program that they will ultimately want to be.

The last thing I'll mention is this issue that we've heard around digital labels, which is like this fascinating thought that, like, the two of us could walk into a store and like, look at the same label and get two different prices. I was recently at a grocery trade show.

I know you're all jealous, but I went around and I talked to the vendors and I said, because they have these electronic labels there. And I said, okay, you know, I get these things are dynamic. What does that actually look like? And I was like. And they're like, oh, well, you know, it's really convenient.

Because that way, like if you have strawberries, like a seasonal item, and you know they're going bad. You can just, like, with a press of a button, like, decrease the price in all your stores kind of thing, right? Or, like, at the end of the day, things like that.

I said, oh, that's really cool, but, like, could you do this in real time? And he was like. Guy, like, looked at me, was like, what do you mean? I'm like. Like, if I'm walking up to something and you know that I really want this box of cereal, like, could you jack up the price?

And he looked at me and he was like, no, and you probably wouldn't want to do that. And I was like, thank you, sir. So I hear this concern, and it does have this, like, dystopian vibe to it, but, like, it's not really something that's currently being deployed, and I don't know if it practically could be.

So I just want to kind of disabuse all of us of any concerns we might have of getting offered different prices for the same item at the same time. With that, I will conclude my testimony.

I was just acknowledging we made time.

Awesome. Thank you both. Ms. Ortega,

I. Yes, and I appreciate that. I think my beginning where I said the rule everyone's going to violate is keeping to time work. Ms. Ortega, you wanted to.

Yeah.

Oh, I'm sorry. Hold on. I didn't ask for opposition of the public. It's not you, it's me. Sorry. Anyone else in the room? In opposition. I didn't mean to ignore you. Apologies.

Appreciate it. Madam Chair Alex Torres, on behalf of the Personal Insurance Federation of California, I appreciate the amends. We'll be reviewing that in touch with the other. Thank you.

Thank you.

Good afternoon, Madam Chair. Member Skyler Wannacott with the California Business Properties Association in opposition.

We got you. This is the opposition line.

Amelia Zamani with the California Travel Association. We align our comments with the chamber. But thank you for the amends. In opposition. Did I say that?

Good afternoon. Chair Members Jose Torres with TechNet and respectable opposition

Marjorie Lee with Samson Advisors on behalf of the California New Car Dealers. In respectful opposition.

Amanda Guadarrama with Cal Broadband. In respectful opposition.

Sabrina Lockhart with the California Attractions and Parks Association. Appreciate the amendment's continued conversation, but we remain opposed. Thank you.

Ryan Allain, on behalf of the California Retailers Association. Respectful opposition, thank you.

I want to know who's paying less to go to Disneyland. Ms. Ortega.

I had a question about the rewards program. So are you using personal information to develop the rewards program? Yes. So how is this, can you explain more?

Yeah. Oh, sorry. Sorry. I should say into the mic, for the record. So like when I've spoken to Members about this, right, it's things like when you sign up for a loyalty rewards program, right? You sign up and then they may say, oh look, you have bought, for example, every year you come around and buy pumpkin pie.

Oh, hey, it's October. Like maybe we want to make sure that, you know, Ms. Ortega comes in and gets that. She doesn't come in. Let's give her a 5% discount on pumpkin pie. Right? Or it may be a store wide issue. Our store hasn't sold as many like detergents, for example, this month.

So now we have an overstock. We should lower that. Let's give a discount for 10% to encourage people to come buy that overstock. Right.

So, and I, and I think many times those kind of discounts, you know, and I say this, you know, with all respect, I think confusion comes up on like, wait, wait, why'd you do that? And that's.

Those are like, I've asked every employer we have saying if there is a, you know, secret here, tell me because I don't know what it is. And they've all said no, that's it. Like, that's how we offer them. But it does under the definitions in the Bill and the CCPA, right.

Even that information is considered personal and is covered. So that's the kind of information we're talking about. I hope that's helpful to the question. Is that responsive?

But if the loyalty program is exempted.

Hold on, It's not a back and forth, sir. Thank you, Ms. Ortega.

Yes, well, you kind of made the point for me and the whole point of this Bill is personal data, information, algorithms. Correct. You know, created for personal shopping. Correct.

So if I'm a mom and I need to go out and buy diapers and the based on my search, because I've searched in, in my area, where can I buy the cheapest diapers because I need them.

And then I get to the store and once I get to the parking lot, those diapers have increased when I got to the store is what I understand.

Do you want to take that or should I? Go ahead. Sorry, I don't want to. You didn't actually ask.

So I guess what I'm trying to say here is, and I hear what you're saying, but based on the work that I've done with the author, to me this seems like modern day redlining where you're using my personal information to come up with the price.

Then, you know, I'm desperate for Those diapers I get to the store and the diapers have increased by whatever amounts of money because you've decided based on the algorithm of my search that I am in desperate need of these diapers and the example I'm showing there, because there's a lot of jargon.

So I try to, I'm trying to break this down for the average viewer who's trying to understand the problem I have with what's happening is that you are using my personal information to price gouge me and maybe and I'll go back to the author to further explain because what I asked you earlier is you are using my personal information and my data to come up with the rewards.

So you're also using my personal information and data to come up with the price based on my need in that moment.

Go ahead, Mr. Ward.

Thank you always for the chair.

I think that what is an important point to be able to hone in on here that is helping us to understand what is sort of a fair practice that we already experienced today, which is that if you join, and I apologize for naming names, but say the Albertsons Save On Club, right, Everybody has the opportunity to be able to join that.

And everybody that's a part of that club is going to be able to receive those coupons and discounts at home and be able to receive those notices to your email, come into my store and it's a sale day and be able to get that dollar off your coffee when you check out. Right?

That is something that is universally available to the General public.

We do want to make sure as this bill and if and as this bill continues to move forward that we are respecting our loyalty programs because hey, what, California doesn't want to save a little bit of money, and that's also very fair and good competition for you and other competitive businesses out there to work on each other about who can offer a good for the lowest service.

And I get that. And I think what we want to be careful as we're having those conversations is exactly the point that Ms. Ortega is making, which is that when we're trying to make sure that what I'm articulating for the need of this Bill is to make sure that somebody is not paying a higher price based on the personal information. Right? You live in a zip code that has a high Latino density.

That information right there shouldn't be informing what price you're paying online for a certain good. But it also should mean that a loyalty program that's out there shouldn't be more available, say to Non Latino zip code holders. Right. Because then the effect, the net effect is the same.

And so the point is, is that the practice of connecting personally identifiable information to the pricing of goods here is what we are trying to. And by the way, it's not just California. Several other states are trying to work on it this legislative year because the FTC is not moving forward on this issue at the moment.

That is what we're trying to accomplish here. And I think that through these amendments today and continued conversation, we'll be able to get there.

Yeah. And I will say, and I'll turn it back over to you In a second, Ms. Ortega. I want to thank the author for his work on the loyalty programs.

I think he and I as parents who grocery shop more than we would like to because the cost of goods these days, fully understand the value of these programs. And he has taken the amendments. It was the intent of the Committee to ensure that they aligned with the CPPA.

Again, I know that was in short order, but this Committee is committed to continued work to make sure that there's not inconsistency in the way the opt in works and other things, because that was not the intent of the amendments.

But I think I've heard the author say that, and the accepting of the amendments today clearly reflect that back. Ms. Ortega, are you done or. Okay. Ms. Pichu Norris and then Mr. Bryan.

All right, oh, I don't think there was a response.

I don't think she asked it. The question at the end, so I didn't want to. Say.

All right, thank you. Thank you, Madam Chair. Okay. So to the author, I will start by actually saying I'm going to use the line from the opposition, which I think this notion of surveillance pricing does seem like this sort of very creepy, dystopian future.

Like none of us really want to be tracked and we certainly don't want to be tracked. And then price gouged on top of that. Right. So I think we all agree on that premise.

I guess I am a little confused about what's already possible for consumers to opt out of because of the CCPA and all of the work we've done as a legislative body and as a business community to implement that. So I guess can you help me understand what's already covered by the CPPA?

So if I opt out at CPPA, you can't use my personal information, period. So you certainly can't use my personal information to determine my price. Is that. Am I understanding that correctly or incorrectly? I guess. What's the.

Well, one of the things is you have, you have to opt out.

But only about like less than 2% of people do opt out because you have to be given the opportunity to opt out. So for the CCPA to apply, you have to opt out. And you have to opt out from multiple parties because there's no one standard way of opting out.

Correct.

It's very hard to opt out. This is saying whether you opt in or opt out, when you're given a price, it should be the price.

I would dispute that it's hard to opt out. I mean, every time I open, every single time I open a website, it says like, you know, and we've made, we have required as a State of California companies to do this.

I can just give you the statistics on how few people..

Please don't interrupt the Assembly Member.

How few people opt out. But I think that actually is illustrative of the fact that there was a very, very like, inefficient use of resources to implement that. We went through this very elaborate process. We required every single technology company in the state to require this.

And at the end of it, like seven people out of eight, 80 million end up opting out. So what that says about what people care about, I don't know. My point is that that exists.

We've gone through an incredibly expensive and you know, it's not just companies that are bearing these costs, like gets passed on to all Californians. We've gone through an incredibly expensive process of establishing that as our standard and that as the way that people avoid this happening. So I guess where are the gaps in that?

Because just saying that not a lot of people are taking advantage of it. That's, that's not a good enough answer. Like, then we should do an awareness campaign about the ccpa, if that's your concern.

Well, one of the things that will change is if, if we had Governor Newsom not veto the legislation last year, that all browsers have to respect an opt out and have to send an opt out and one opt out works for everything. And that was something the Governor vetoed.

I don't know if I think the bill's coming back to him, but it does it. Unfortunately, people are unprotected and this Bill creates another round of protection. I do think one day there will be an app that you hit and I opt out, and I opt out and it sends a signal everyone has to respect it.

But getting there amongst all the opposition and getting the right regulation to do that is very difficult. There is, there are some regulations in work by the Privacy agency to do that. And it is a slow road. I agree with you. And it's very frustrating. Very frustrating.

Okay, so sorry, that's a different question than what are the gaps if I opt out of CPA. Yes. Then I'm asking the author. So what are the gaps that exist and your Bill is trying to fill?

Well, if you opt out of the CPA, that is indistinguishable from what we are trying to solve for here in the Bill, which is that if there is a linkage between the pricing as a, as a, as a marketer, the pricing of a good and what this individual A or individual B is going to pay based on their personally identifiable information, that's what's defined within this Bill and would become unlawful.

So that is irrespective of whether somebody's opt in or opt out of CCPA. But we are trying to make sure that there is no conflict with CCPA.

And if I may just clarify a little bit on the CCPA. So what Assembly member Petrie Norris is pointing to is cookies, which, yes, every time you go to a website, you can opt out of that, that you all see and know. And it's, you know, annoying to a lot of people.

I think Samuel Wilson told me she doesn't ever opt out because she doesn't like the ads she gets. But, but then there is the ability to opt out of the use of personal information beyond the cookies. You may opt out of the use of sensitive personal information, and that is what you would imagine it to be.

But for example, your credit card spending is not in that list. So there is information under the CCPA that remains available to these retailers to use or not use as they want. And so I guess the example that was used would still apply. So I just want to be clear about what the CCPA does and doesn't protect.

Now, that doesn't relate to whether you support this. I wanted to make sure we all understood that.

Thank you. That is helpful. And also, for example, the, I think using an algorithm based on some of that information that is obtainable out there, not covered by CCPA. Right. That algorithm can then start to try to create some kind of a profile of your identity.

And it's that profile that, that algorithmic profile that would start to create differential pricing based on who you are. Right. That's what we're trying to sort of the, the loophole that we're trying to close here. Okay.

And then I know, and I had shared with you before the hearing, I think that whatever solution we come up with, I think it is actually going to be very confusing to the average Californian, even for those who support and like it.

And that's why I'm very, I am concerned about the provision of the that it creates a new private right of action. Because I think that particularly given how confusing this is, it's going to create a firestorm of lawsuits and kind of unfounded and inaccurate accusations that people are still using this information. So that's a piece.

I know it's not the jurisdiction of this Committee, so I know we won't get into it.

But I, as I shared with you, like that's a concern that I have particularly for, for something like this that's so broad based and frankly very confusing, you know, not just the Members of this Committee, but certainly I imagine, you know, to the average Californian.

So as I shared with you, I can't support it at this time, but I know that you're going to continue to work at it and so would certainly take another look as it comes to the floor.

Thank you. We're going to go to the Vice Chair and then to Mr. Bryan, if that's okay. Mr. Bryan, sorry, I keep, I don't know.

Well, I just want to follow up on Assembly Member Petri Norris question just now on the private right of action. Is it subject to the private right of action? I know there's civil penalties, but is so anybody can bring a lawsuit, any individual can bring a lawsuit and then you have class action attorneys putting lawsuits together.

So the short answer is yes, because as somebody is realizing that they are becoming a victim of this differential pricing, we want that person to be able to be made whole. Ultimately, we want the practice to stop and we hope it's a deterrent enough that it doesn't happen in the first place.

But should somebody be using your information to make sure that you're paying something different than say a male counterpart is paying and that frustrates you because you should be paying the same price and they were using your information. I would want you to be able to have that, to be able to have restitution.

Okay, just back to the original, original premise. We talked a few moments ago about opting in, opting out. Yes, there is a small percentage, but the law was just passed in 2018 or so. I mean, changing behavior takes a period of time.

I opt in and opt out and I often, depending on the retailer, sometimes I opt in because I want them to be successful in understanding the marketing needs of their customers. So I sometimes do yes, sometimes I do. No.

But listening to the discussion on whether I going to buy diapers at store X and then I get there and the price has changed. How long will you continue to buy diapers at store X?

If you feel that your prices are being variable, are variable be depending on you act with your feet, you go someplace else, you buy someplace else, you shop someplace else. If you feel there's discriminatory or predatory practices, that's what the marketplace does.

But we're assuming that there are bad systems out there trying to confuse people into dragging them into lower prices and then suddenly bait and switch, there are higher prices. I would personally think that this is getting into a nanny state Watchdog state that is concerning to me.

I would like to see the effects of the Consumer Privacy act continue and the habitual users use the opt in, opt out. And also there are many websites that will too many ads, that's another story.

Just ads that are on the Internet that tell you you could get a lower price if you do this or you could get a lower price. I don't object to that. I object to the proliferation of it, but I don't object to someone gaging.

And we all know that through the marketing data that if I'm suddenly looking at hiking boots, I'm suddenly going to get ads, a plethora of ads on my website or my webpage about oh, you're into hiking boots and I just wanted one pair, you know. So that is.

That is really considered a benefit of online marketing and user information. The plethora of information that's available to the consumer if they feel that they're being overpriced, there's Yelp, there's social media that don't go to store X. They just charge me too much for diapers.

We have more opportunities than we ever have as consumers to broadcast bad behavior by anyone you feel is giving you bad behavior, whether it's restaurants or stores or gasoline companies. And I think it's effective. I think social media commentary is very effective.

And I would gage that that business who wants to continue to be successful will, if it is doing that, it will realize that there are other ways to market their products and be successful.

Well, thank you madam. I share. I would agree with you that there is definitely recourse but often that people are using on their own to put out there and certainly make it their own decision decisions as an individual consumer, whether or not they're going to go patron that store again or not agree. And often they don't.

They're frustrated they're going to go somewhere else. But most of the time is that what we're experiencing today? People don't know what they don't know. They don't know this is happening to them at all. Until you realize that you tried to shop for those boots, those hiking boots and then you live in an Orange County.

You ask a friend, how often do you call that friend up in Fresno and say, hey, would you go online right now and check this pair of hiking boots and find out that friend could actually buy from the exact same website at the exact same time for $40 cheaper?

Because we all presume that people from Orange County can pay more.

You know what I call that store and I say, I just found out that somebody in Central...

You have the freedom to do that is the point. Trying to make sure that we are stopping the practice in the first place. And to say that this is dystopian and this is not happening or this is kind of, you know, not, not really out there is not supported by evidence that already is stuck right now at the FTC where McKinsey&Co.

Is actively coaching corporations to say by using these algorithmic determinations, you can pay to, you can make 2%-7% more in profit. And so there's an encouragement right now for industry to go this way.

And this is the inflection point that we're at here in California saying we absolutely want to make sure that we're adhering to long standing practice. That consumers should pay a fair and similar price is the way that we've always done business. The incorporation of our personally identifying information into individualized pricing crosses a line, jumps the shark. And that's what this Bill aims to solve.

But there's always been variable pricing information for discounted programs or loyalty programs. You get a lower rate on your airline ticket if you have X number of thousands of miles. They may give you a little bit break on different seats.

That doesn't change under this Bill.

Well, anyway, I, maybe this is the early days. Let the FTC decide what to do, I suppose. But I think this is just one more constraint on doing business in California that creates problems for all, thank you.

Thank you, Madam Vice Chair. Moving to Mr. Bryan. Sorry for the delay.

No, thank you, Madam Chair. I want to thank the author and Mr. Court. Always good to see you. Yeah, I think the author just addressed the fact that we're not changing discounts or variable pricing. And we know that the market already has versioning.

You can buy the Toyota Corolla or The Lexus whatever, at different price points, even though they're essentially the same car, because they are marketed at different points. I think the purpose of this Bill is you can't sell the same Corolla to two different people for two different prices because of characteristics that you find out about them.

I see this all the time in my district because I represent a very interesting district, if you know, Los Angeles. I represent Crenshaw and Slauson, where Nipsey Hussle grew up, and Beverly Hills, where the Beverly center is.

And I can tell you that the cost of Internet, which is why I'm glad to see my friends Cal Broadband here is different in South LA versus Beverly Hills, but it's actually unique and different.