Latest / CLEs You Actually Want to Hear / A Fireside Chat with the FTC’s Former Consumer Protection Director with Samuel A.A. Levine, Christina Tusan
Transcript
- 0:03Welcome to the FTC Privacy and AI Enforcement
- 0:05Trends, a fireside chat with the FTC's former
- 0:08consumer protection director. I'm Christina Toussin.
- 0:11I'm the chair of the Beverly Hills Bar Association's
- 0:14Privacy and Cybersecurity Law section. And I'm
- 0:17delighted to introduce today's speaker, Sam Levine.
- 0:20Sam previously served as a director of the FTC's
- 0:23Bureau of Consumer Protection, where he oversaw
- 0:25enforcement, rulemaking, and policy work. across
- 0:29a wide range of areas, including privacy, data
- 0:32security, marketing, financial services, digital
- 0:35advertising, consumer reporting, algorithmic
- 0:38decision -making, small business financing, just
- 0:41to name a few. Before his appointment as director,
- 0:43he served as an attorney advisor to Commissioner
- 0:45Rohit Chopra and as staff attorney in the Midwest
- 0:48Regional Office. Prior to joining the FTC, Mr.
- 0:52Levine worked for the Illinois Attorney General
- 0:54where he prosecuted predatory for -profit colleges
- 0:57and participated in rulemaking and other policy
- 1:00initiatives to promote affordability and accountability
- 1:03in higher education. And now let's get started.
- 1:08So Sam, we're really happy to have you here with
- 1:10us. You and I worked here at the FTC, so I know
- 1:13some of your roles there, obviously, but let's
- 1:15talk about your background and roles at the FTC
- 1:18and how those changed over time. Sure. Well,
- 1:21first, it's really wonderful to be here. I actually
- 1:25got my start in consumer protection during the
- 1:27financial crisis. I was in law school at the
- 1:30time. I got involved in a legal clinic where
- 1:33I was representing people in foreclosure. That
- 1:36made me really passionate about using consumer
- 1:39protection tools to advance economic justice.
- 1:42It also made me really angry, frankly, at what
- 1:44little the government was doing both to prevent
- 1:47the mortgage meltdown, help homeowners through
- 1:51that mortgage meltdown, and I know Californians
- 1:53were very hard hit, and hold accountable the
- 1:56executives that I think caused the mortgage meltdowns.
- 1:58That gave me, put a real fire in me around consumer
- 2:01protection. I took that fire to the Illinois
- 2:04AG, where as you said in the intro, I really
- 2:08focused on for -profit colleges. I got to meet
- 2:10Rohit Chopra at that point, who was student loan
- 2:12ombudsman. I then wanted to work with the Federal
- 2:17Consumer Protection Agency, the FTC, so I started
- 2:20in the Chicago office there. I moved over in
- 2:242018 to advise Rohit Chopra, who was one of the
- 2:28Democratic appointees on the commission during
- 2:31the first Trump administration until about two
- 2:33months ago. The FTC was a bipartisan commission.
- 2:39Commissioner Chopra sat on the commission until
- 2:412021. And then in 2021, when Lena Khan, who had
- 2:46been a fellow in our office just a couple of
- 2:48years before, became chair, she named me head
- 2:51of the Bureau of Consumer Protection. And for
- 2:53those who don't know, the FTC is more or less
- 2:57comprised of three bureaus. Bureau of Economics,
- 3:00which does economic analysis across the agency's
- 3:03work. Bureau of Competition, which enforces the
- 3:05competition laws in conjunction with the Department
- 3:08of Justice. And the Bureau of Consumer Protection,
- 3:10which enforces consumer protection laws. The
- 3:13FTC Act, of course, which prohibits unfair deceptive
- 3:16practices, but also a host of other laws. Fair
- 3:19Credit Reporting Act, Children's Online Privacy
- 3:21Protection Act, many other laws. And BCP, in
- 3:25addition to doing enforcement both in federal
- 3:28court and And it's an in -house court that also
- 3:30does rulemaking, which I know is something we're
- 3:32going to talk about. So really broad portfolio
- 3:34at BCP. It's a team. It's about 440 people. That
- 3:39sounds really big. That's actually a very small
- 3:41group responsible for rooting out unfair deceptive
- 3:46practices throughout just about the entire American
- 3:49economy. So I really like to say the FTC punches
- 3:54above its weight. It's a small and mighty team.
- 3:58As you know well. I can't agree with you more.
- 4:01Having worked there, really amazing attorneys
- 4:03doing some really, really important work. Definitely
- 4:06small but mightiest and certainly my description.
- 4:09So when you were director, what were some of
- 4:11the priorities in terms of the types of cases
- 4:13the FTC was bringing in the privacy and AI space
- 4:16as you sort of shaped that as a director, kind
- 4:19of looking into what priorities you were looking
- 4:22at when you were there? Well, when I became director
- 4:25in 2021, AI was not the hot topic that it is
- 4:28today. And one of the things I'm proud of is
- 4:32I think we were a little ahead of the curve there
- 4:34in the summer of I think it was the summer of
- 4:362022. We issued a major report on the uses of
- 4:40AI around how AI can and cannot be used to challenge
- 4:46certain consumer abuses like fake reviews and
- 4:49bias and others. And when the generative AI hit
- 4:53in, what was that, late 2022, I think was when
- 4:56ChatGPT rolled out, we were able to move very
- 4:59quickly to start bringing a whole host of cases.
- 5:03We started by laying out really comprehensive
- 5:05and award -winning business guidance and how
- 5:08our laws apply to the use of AI. And then we
- 5:11moved pretty quickly to bring in cases. So we
- 5:13brought cases around unsubstantiated claims about
- 5:16the impact of AI, AI decision -making systems
- 5:19that result in errors and discrimination, all
- 5:25sorts of cases around companies that were used.
- 5:28promising people could get rich using AI tools.
- 5:32We shut those companies down. And we also did
- 5:35some rulemaking, which I can talk about to try
- 5:37to get at AI fraud, which is a major concern
- 5:40and a growing concern. On the on the on the privacy
- 5:45side. And again, these are related, obviously,
- 5:48but on the privacy side. Very early in my tenure,
- 5:52I said that the goal of the privacy enforcement
- 5:55work of the FTC would be to move away from notice
- 5:58and choice, which is this notion probably most
- 6:01of your listeners know. It's basically the privacy
- 6:04regime we've lived under for decades, except
- 6:07in California, where people are essentially expected
- 6:12to protect their own privacy online by reading
- 6:14lengthy terms of service and privacy policies
- 6:17and deciding whether or not they want to use
- 6:19the service. It's a total fiction. It's been
- 6:22a major failure. Companies can decide at will
- 6:25whether they want to change these policies. It's
- 6:28completely unrealistic for consumers to expect
- 6:30themselves to protect themselves. So what I said
- 6:33early on is this regime has been a failure. Be
- 6:38honest that this has been a failure and we have
- 6:40to use other tools and the biggest tool that
- 6:43I Advocated using at the beginning and I think
- 6:45we delivered on this over the last four years
- 6:48was unfairness You know, historically, the FTC
- 6:52focused its privacy enforcement on deception.
- 6:56Are companies misleading people about how their
- 6:58data is being used? That's certainly something
- 7:01we continue to do under my tenure and under Chair
- 7:04Khan's tenure. But we shifted a lot to bring
- 7:07in a lot more unfairness actions where we said
- 7:10whether your privacy practices were disclosed
- 7:13or not. They were substantively harmful. They
- 7:16put people at risk of substantial injury. That's
- 7:18the test for unfairness under the FTC Act. And
- 7:21we brought a whole series of cases using our
- 7:23unfairness authority to not just require that
- 7:26data sharing or data use be better disclosed,
- 7:29but to actually restrict the ways in which companies
- 7:32can use people's data, especially their sensitive
- 7:34health data, their location data, browsing data,
- 7:37kids' data, et cetera. So the big theme of my
- 7:40tenure was moving toward substantive protections,
- 7:44data minimization, as opposed to notice and choice.
- 7:47And of course, in the background, as we were
- 7:50advancing that agenda, we saw a whole lot of
- 7:52states. California led with this early on, but
- 7:58we saw a lot of states over the last few years
- 8:00pass their own privacy laws, most of them grounded
- 8:03in data minimization principles. So I really
- 8:06saw the FTC as trying to lead a whole of government,
- 8:09not just federal, but also state government effort
- 8:11to move away from notice and choice and toward
- 8:14real safeguards for people's privacy and personal
- 8:17information. Well Sam, that's a really important
- 8:20point because I know one of the challenges in
- 8:23working in this space is, you know, somebody
- 8:25makes a deceptive claim. And they're told that's
- 8:28a deceptive claim, and then they can go forward
- 8:30and simply change the claim and the small terms
- 8:32and conditions, and then the consumer is still
- 8:34no better off, right? The deceptive practice
- 8:36may continue. So I think it's really important
- 8:39that distinction you made and really moving towards
- 8:41the unfairness analysis and being able to sort
- 8:44of really make a change in that space. And that's
- 8:46just such a really important shift and really
- 8:49great that the FTC moved in that direction, because
- 8:52I think we all are familiar with the very large
- 8:54terms and conditions that maybe Nobody besides
- 8:57you and me, and well, sometimes my children,
- 9:00I forced to read, but I think most - I don't
- 9:02read them. I don't have time. I read them for
- 9:06my own purposes. Yeah, correct. Fair enough.
- 9:10Well, let's talk a little bit about kind of the
- 9:14evolution of the FTC's privacy and AI space.
- 9:17You talked a bit about that before. I know an
- 9:20FTC report has talked about since 1999, there've
- 9:23been 97 privacy cases and 169 telemarketing sales
- 9:28rule and can't spam cases, really a lot of activity
- 9:31over the years there. And of course, affecting
- 9:34millions of consumers, getting really significant
- 9:36restitution and changing practices, which is
- 9:39great. And you talked a little bit about how
- 9:41it's changed in recent years, but can you give
- 9:44a more of a sense as to how much it shifted in
- 9:47terms of privacy and AI really being more of
- 9:50a focus than maybe it was, even though there
- 9:52were a number of cases over the past years? Yeah,
- 9:56I think the history here is quite interesting.
- 9:59I'll try not to drone on too long about it, but
- 10:03when the internet... emerged when Al Gore invented
- 10:07it, when it emerged and personal computers in
- 10:11the 90s, the FTC was very ahead of the curve.
- 10:16There were a whole series of reports. I'm sure
- 10:18I know you know this, Christina, that the FTC
- 10:20issued in those days sounding the alarm around
- 10:23kids' privacy. It's led to passage of COPPA,
- 10:26Children's Online Privacy Protection Act, around
- 10:29websites that were tracking people across the
- 10:32web, around what information consumers might
- 10:35be turning over to scammers, thinking that they're
- 10:38turning over to legitimate companies. And I think
- 10:43the reason this work fell to the FTC is because
- 10:47there was no comprehensive privacy law then and
- 10:51there's no comprehensive privacy law now. There's
- 10:54no privacy enforcer then, nor is there today.
- 10:58So sort of by default, it fell to the FTC. And
- 11:00now I should say there is some history for why
- 11:02the FTC became the de facto chief privacy enforcer.
- 11:06We had long, I still say we, I don't speak for
- 11:10the FTC, never did, I certainly don't now. We
- 11:13had long enforced the Fair Credit Reporting Act,
- 11:16going back to when it was passed in, I think
- 11:18it was 1970. We had long enforced Gramm -Leach
- 11:23-Bliley, which mostly was a deregulatory bill,
- 11:26but also provided certain privacy protections
- 11:28for consumers of financial services. And of course,
- 11:32the unfair deceptive practices authority is quite
- 11:35broad and can be adapted to addressing privacy
- 11:38harms. So that's how the FTC essentially became
- 11:41the de facto privacy regulator. And in the 90s,
- 11:43as they said, the agency was really prescient
- 11:45in sounding the alarm around practices. And the
- 11:48hope then that the FTC expressed was that companies
- 11:52would take self -regulation seriously. and would
- 11:56develop their own frameworks to protect people's
- 11:58privacy. That did not work. Year after year,
- 12:02the FTC expressed frustration that companies
- 12:05were not doing so. And in 2000, this is still
- 12:08under the Clinton administration, the commission
- 12:10called on Congress to pass comprehensive privacy
- 12:12legislation. That was 25 years ago. I think they
- 12:16were really ahead of their time on that. However,
- 12:20there was an election that November. The administration
- 12:24changed hands the following year. And when the
- 12:28new FTC administration came in in 2001, one of
- 12:32the first things they did is they said, we're
- 12:34changing our mind, Congress should not pass privacy
- 12:37legislation. We do think that self -regulation
- 12:41can work. And we're really going to focus on
- 12:44offline harms as opposed to online harms. We're
- 12:47going to focus on do not call violations, unwanted
- 12:50telemarketing, identity theft. That was a pretty
- 12:55narrow focus. Over time, I think the FCC definitely
- 12:59broadened its enforcement and started bringing
- 13:01unfairness cases to challenge reckless data security
- 13:05practices. That happened, I think the first case
- 13:07was in 2005. It started bringing deception cases
- 13:11during that same period around companies that
- 13:14misled people about their privacy policy. And
- 13:17over the years to come, you saw the FTC, you
- 13:20know, wax and wane depending on the administration.
- 13:22But certainly, and you know, that's Christina,
- 13:24during the Obama administration saw the FTC very
- 13:26active and trying to get Congress to pass privacy
- 13:29legislation and bringing privacy cases of its
- 13:33own. I think the big shift that we tried to effectuate
- 13:37under Chair Khan over the last four years was
- 13:41moving beyond deception. As I said, the FDZ had
- 13:45used unfairness to challenge data security practices,
- 13:49the FTC hadn't really used unfairness to challenge
- 13:52privacy practices that maybe were disclosed,
- 13:56but that still harmed consumers. So the big change,
- 13:59as I alluded to earlier, that we tried to make
- 14:01over the last four years was recognizing that
- 14:05in a world where everything we do is often mediated
- 14:09through digital services, whether it's connecting
- 14:11with our friends, keeping track of our fitness,
- 14:13keeping track of our health, We can't rely on
- 14:16consumers just read ever more policies. We need
- 14:20some basic bright line guardrails on how our
- 14:22information can be used. If I'm going to go to
- 14:25our racks to get a discount on a prescription.
- 14:29It shouldn't be the case that good or rats can
- 14:31then sell information or share information on
- 14:34what prescriptions I'm using with Facebook. And
- 14:37the fact that that might be disclosed somewhere
- 14:39does not make it OK. So that was the position
- 14:42that we moved toward over the last four years.
- 14:44And I know we're going to talk about this. We
- 14:46brought a whole series of cases making clear
- 14:49that the FTC's enforcement, the FTC's unfairness
- 14:53prohibition. actually restricts the kind of sharing
- 14:57of sensitive data that companies too often engage
- 15:01in without regard to how it could harm consumers.
- 15:05Yeah, that's great. That's a really, really important
- 15:07background in terms of how things have progressed
- 15:10there. So let's switch. We've talked about the
- 15:12privacy space a bit and let's switch to the AI
- 15:15space, which as we've discussed, FTC did some
- 15:19early analysis on, but let's talk about some
- 15:22of the cases that the FTC has done in the AI
- 15:24space. And maybe we can start with talking about
- 15:27the Rite Aid AI bias case and kind of give the
- 15:30listeners a sense as to what that case was about
- 15:33and maybe some lessons learned from that. Rite
- 15:36Aid was a really important case. So I don't recall
- 15:39whether Rite Aid is a big chain in California.
- 15:41Maybe you can tell me. Yes, that's right. But
- 15:47Rite Aid, of course, is a chain of retail pharmacies.
- 15:52They deployed an AI facial recognition system
- 15:57to try to deter shoplifting. But the system had
- 16:03serious, serious flaws, serious, serious accuracy
- 16:06issues, and Rite Aid's employees, as the FTC
- 16:09alleged in its complaint, excuse me, were very
- 16:12poorly trained on how to use the system. both
- 16:16how to upload photos of suspicious persons and
- 16:21how to use the system's recommendations to identify
- 16:24suspicious persons who were in stores. And this
- 16:27particularly impacted people of color and women
- 16:31for whom the error rate was much higher. And
- 16:33what we allege in the complaint is that it got
- 16:35so bad that there were young mothers and their
- 16:40daughters wrongfully detained in Rite Aid stores
- 16:42accused of shoplifting because they were falsely
- 16:46matched with someone else in some other state
- 16:48who had actually been accused, who had actually
- 16:53either been engaged in shoplifting or... been
- 16:55tagged as engaging in shoplifting. So you see
- 16:59here kind of the real -world impact. People talk
- 17:01about AI like it's an abstraction. This is when
- 17:04the rubber meets the robe. This is when this
- 17:06AI facial recognition system was deployed so
- 17:09recklessly and the system was so sloppy and error
- 17:13-prone that it resulted in real, very tangible
- 17:16harm for people in these stores who were innocent,
- 17:20but they were wrongfully detained and humiliated.
- 17:24So the FTC alleged the use of this system without
- 17:28taking adequate safeguards was unfair under the
- 17:33FTC Act. And we actually prohibited Rite Aid
- 17:37from deploying AI facial recognition for five
- 17:39years and required that if Rite Aid does deploy
- 17:44any kind of biometric surveillance system in
- 17:46its stores. It needs to do detailed and regular
- 17:49auditing to make sure that the system is accurate,
- 17:52to make sure that the system does not result
- 17:54in bias. Now, I think this was a really important
- 17:57case for a number of reasons. First, I think
- 17:59we stopped a really egregious harm that especially
- 18:04women and people of color who were going into
- 18:06these stores were facing. Second, I think we
- 18:09sent a really strong signal to the market that
- 18:13companies can't just purchase AI systems and
- 18:17install them. They really need to be careful
- 18:20and to ensure that these systems are accurate,
- 18:23to ensure that employees are trained in how to
- 18:25properly use them, to ensure that the regular
- 18:27being audited for accuracy and bias. And I think
- 18:31it was also important in terms of the development
- 18:33of FTC law. If we had brought this case, I don't
- 18:37know that we would have brought this case a decade
- 18:39ago. But if we had, it might have been, for example,
- 18:43a deception theory. Maybe Rite Aid should have
- 18:48told people that they were using these systems.
- 18:50But that is not the theory we brought in the
- 18:52case. What we said was substantively unfair for
- 18:55Rite Aid to do what Rite Aid did. And that's
- 18:57how we were able to get the strong remedy. That
- 19:00resulted so I think you know right it was an
- 19:02important case in its own right I think it also
- 19:05sent based on what I've certainly heard from
- 19:06any defense lawyers a really important signal
- 19:09to other market participants and by the way I've
- 19:12heard for many compliance lawyers that right
- 19:14aid was helpful in Their counseling their clients
- 19:18that they need to be really careful when they
- 19:20deploy these technologies again. I should have
- 19:22said at the top I am not anti -AI in the least.
- 19:28I use AI tools for all sorts of things. I think
- 19:31it offers a lot of promise. But what I believe
- 19:34fundamentally is that the laws that govern every
- 19:37other corner of the market, there's no AI exemption.
- 19:42As with the use of any tool, companies that use
- 19:45AI tools need to ensure it's not harming people,
- 19:48need to ensure it's not discriminating against
- 19:50people, need to ensure they're not making false
- 19:52claims about the abilities of these tools. So
- 19:55I really just see this as applying the law to
- 19:57a new technology, but just something the FTC
- 19:59has been doing for its whole existence. In the
- 20:03same way you wouldn't have an employee that you
- 20:05weren't monitoring who was making statements
- 20:07or doing things that weren't approved, the AI
- 20:09models need to really be monitored in that. Exactly.
- 20:13Yeah, absolutely. In the AI space, talking about
- 20:16some other cases, can you talk about the Ring
- 20:18case and what the circumstances were of that
- 20:21case and maybe lessons learned, is it there as
- 20:23well? Well, Ring was a home security camera,
- 20:26Amazon. purchased Ring at some point, I don't
- 20:29remember when. And what we allege, among other
- 20:32things, was that Ring employees were spying on
- 20:36people, were spying on people in their homes
- 20:40using this technology, thanks to really sloppy
- 20:43back -end protocols within the company. We required
- 20:47Ring, I can't remember how much... They had to
- 20:49pay. I want to say it was tens of millions of
- 20:52dollars. I'm sorry I'm blanking on that. This
- 20:55is why sometimes AI is better than humans. And
- 21:00we brought that case in conjunction with another
- 21:03case against Amazon, Amazon Alexa. And I think
- 21:06it's important to view these cases in tandem.
- 21:08We announced them the same day, because in that
- 21:10case, what we alleged is that Amazon was retaining
- 21:13children's voices indefinitely, which violated
- 21:18Children's Online Privacy Protection Act, and
- 21:20using those voices to train their AI model. And
- 21:25the message that we wanted to get out through
- 21:27that case and that we reinforced in our recent
- 21:30updates to the COPPA rule is that COPPA prohibits
- 21:36retaining people's information indefinitely.
- 21:38And there's no exception to that because you
- 21:41intend to train AI models. So in that case, again,
- 21:46was brought with the Department of Justice. They
- 21:48paid a civil penalty of tens of millions of dollars.
- 21:52And I think the most important message there
- 21:53was that COPPA's prohibition on indefinite retention
- 21:58applies. AI does not create an exception. Great.
- 22:02That's a really important point. It's not just
- 22:05how you're collecting the data, it's what you
- 22:06do with it once you have it. There are lots of
- 22:08rules around not over collecting and what you
- 22:11do after you have it, and I think that's really
- 22:13important. You were right, it was 5 .8 million
- 22:16in the ring case, so it was millions and millions
- 22:19of dollars. Then I think Alexa was 20 something.
- 22:23I was thinking of them together. I thought it
- 22:26added up to $30 million, but again, go ask ChatGBT,
- 22:30it can tell you. Yes, absolutely. Absolutely.
- 22:33So now that that's great. So I mean, those are
- 22:35really important cases. And obviously, the FTC
- 22:38is moving quickly in a space that I think a lot
- 22:41of a lot of companies and individuals are trying
- 22:43to to sort of catch up on. You know, what advice
- 22:47do you have for companies who are looking to
- 22:48incorporate AI into their businesses? We talked
- 22:51a little bit about some sort of sort of lessons
- 22:53learned here. But but any more general advice
- 22:55as people sort of are looking at the newest trends
- 22:58and thinking they can they can sort of move their
- 23:01business along more quickly and improve things,
- 23:03what should they be considering as they do that
- 23:05evaluation? I mean, so there's no shortage of
- 23:08ways that AI can be useful. And there's no shortage,
- 23:12I'm sorry to say, of companies to get in trouble
- 23:17due to the use of AI. So maybe I'll focus on
- 23:19two in particular in addition to those I already
- 23:23mentioned. I'll say three. First, I understand
- 23:28we want to see a vibrant competitive market for
- 23:31developing AI tools. But that can lead to a temptation
- 23:36to overstate the ability of these tools. And
- 23:38the FTC has brought a whole series of cases.
- 23:40And by the way, some of those cases are continuing
- 23:43in this administration, which I think is notable,
- 23:46around claims companies are making about their
- 23:48AI tools. So we brought a case against a firm
- 23:50that claimed their facial recognition was bias
- 23:53-free. It wasn't. We alleged that was deceptive.
- 23:57We brought a case. against a company called Do
- 24:00Not Pay. I believe the California Bar Association
- 24:02also had a run in with them, which was an AI
- 24:06lawyer. I don't have any problem. Some of the
- 24:09lawyers in the crowd might try to throw tomatoes
- 24:12at me, but I don't have any problem with people
- 24:15consulting. AI around possible legal problems.
- 24:20But you need to talk to your human too. What
- 24:23Do Not Pay was saying was that you don't need
- 24:25a lawyer as the FTC alleged in its complaint.
- 24:30This can be your AI lawyer. They were overstating
- 24:33the ability of their technology to actually act
- 24:36as an attorney. We allege that was deceptive.
- 24:39We made clear by the way that We're not against
- 24:42using AI, we're against making false claims about
- 24:45AI. So if you're selling an AI tool, or if you're
- 24:48using an AI tool and you want to represent how
- 24:51you're using it to consumers, make sure your
- 24:56claims are backed up. I'll give one more example
- 24:58in that space of a case we brought with the LA
- 25:01District Attorney actually against a company
- 25:03called NGL. It was an anonymous messaging app.
- 25:05I won't get into all of the rather gory allegations,
- 25:08but they claimed that they had an AI I think
- 25:11it was content moderation system to keep kids
- 25:13and teens safe on the platform. Good reason to
- 25:18allege that that claim was not substantiated.
- 25:22New Jersey just brought a case. You're seeing
- 25:24a lot of AI safety claims now. Those need to
- 25:27be backed up too. Two other quick pieces of advice
- 25:30I would give. One, companies are increasingly
- 25:35turning, I've seen this in my personal life,
- 25:38to AI chatbots. to provide customer service.
- 25:44My advice there is you're responsible for what
- 25:47your chatbot spits out. I have heard about examples
- 25:50of chatbots making deals with consumers that
- 25:54maybe the company did not like. My view, other
- 26:01enforcers might take a different view, but my
- 26:03view is that you are clearly responsible for
- 26:06this. Chatbot is essentially acting as your agent.
- 26:10You are responsible for what that chatbot represents
- 26:14to consumers. And on the flip side, if I ask
- 26:21a chatbot, are you going to retain this chat
- 26:25when you delete data? Do you sell my data? Do
- 26:29you share my data? And it lies to you. I think
- 26:33the clearly right answer under consumer protection
- 26:36law is that the company responsible for that
- 26:40chat bot is responsible for that. If those are
- 26:42false claims, the company's on the hook. So companies
- 26:45should really be very, again, it's not about...
- 26:51being anti -AI, it's about making sure you realize
- 26:54that if you're gonna get rid of the human, just
- 26:55as you're gonna be responsible for what your
- 26:58human customer service agent says to a consumer,
- 27:00you're gonna be responsible for an AI customer
- 27:03service agent. Last point I would make, because
- 27:06I know it's a hot issue in Sacramento and it's
- 27:08a hot issue in state capitals all over the country.
- 27:12Surveillance pricing. People also know it as
- 27:15personalized pricing. There are a lot of, at
- 27:19this point, it's mostly consultancies out there
- 27:22offering to help companies instead of sort of
- 27:25having one price for one product, have one price
- 27:28for one person. To personalize prices for Christina,
- 27:32for Sam, based on... My location, my demographics,
- 27:37my purchase history, my browsing history, you
- 27:40name it, all the data that's out there and all
- 27:42of us can be used to set an individualized price.
- 27:46And these companies are promising, these consultancies
- 27:48are promising that this can be profit maximizing
- 27:50for sellers. There are bills in a number of states
- 27:55to prohibit surveillance pricing altogether.
- 27:59We're happy to talk about that. But the fact
- 28:03that those bills exist does not mean that it's
- 28:05lawful in all instances today. I think there
- 28:08are significant legal risks around using these
- 28:12perhaps third -party pricing tools. You run the
- 28:15risk of discriminating against people. You run
- 28:20the risk of price gouging. Let's say the system
- 28:24during a natural disaster jacks up prices in
- 28:29different parts of the country that are being
- 28:31affected. And you run the risk of really pissing
- 28:34off your customers. Now, that's not illegal,
- 28:38but that is something that should be. And I should
- 28:40add one more risk. We've seen a lot of enforcement
- 28:43both in the private bar. and buy enforcers in
- 28:46many states and municipalities have laws, so
- 28:49-called fake discount laws. Companies that are
- 28:52claiming this is 20 % off, but it's not really
- 28:5520 % off. Maybe a product is always 20 % off,
- 28:59or it's 20 % off a price that isn't actually
- 29:01charged to anyone. If you're using algorithmic
- 29:04pricing and you're telling people they're getting
- 29:06a discount and they're not actually getting such
- 29:09a discount, it's instead of a personalized price
- 29:11tied to their demographics, that's another way
- 29:13you can run a foul of the law. Oh, and one more
- 29:16legal risk. The longer I go, the more I think
- 29:19about this. If you are using these third party
- 29:21tools, there is a real risk that they could be
- 29:24facilitating collusion. The Department of Justice
- 29:27sued Wheelpage and a bunch of large institutional
- 29:31landlords for using algorithmic rent setting
- 29:35in a way that that was facilitating collusion
- 29:38among these landlords. That same risk exists
- 29:41in other industries too. So I would just urge
- 29:44companies to really proceed with caution if they're
- 29:47going to engage in personalized pricing. Certainly,
- 29:51it can piss off. I should use less salty language.
- 29:54Certainly, it can frustrate your customers. But
- 29:57it could also run afoul of a number of laws that
- 29:59are already on the books. And couldn't there
- 30:02be one other way, which is maybe you have a national
- 30:05advertising campaign for a certain price point,
- 30:07right? But then that's not really available to,
- 30:10you know, to Christine or Sam because they've
- 30:13decided we live in an area where they think they
- 30:14can charge us more, right? So I think there can
- 30:17be another yet one more reason, which is if you're
- 30:20advertising, we provide this hotel room or this
- 30:23property or this product at this price, but somehow.
- 30:27because they've decided that you or I will pay
- 30:30more, they'll charge us more. I think there could
- 30:33be another piece of it in the advertising side
- 30:36of when consumers find out, wait, just because
- 30:38I live in this neighborhood or I have this background,
- 30:41I'm getting charged more. Aside from potential
- 30:43discrimination, there could also be false advertising
- 30:46maybe in connection with those. Is that a possibility
- 30:48as well that you can see? It's an excellent point.
- 30:51You have some cities and states. I think California
- 30:53has a law like this that says, You can't charge
- 30:56someone more than the price you advertise. That
- 31:00is a serious risk, too. I really feel as I have
- 31:06not been a plaintiff's lawyer, but if I were
- 31:09and I had a client who believes they were harmed
- 31:14by personalized pricing, it's a law school issue
- 31:17spotter, all the issues that can emerge. And
- 31:20again, that. That doesn't mean state law, additional
- 31:25protections are not necessary, nor does it mean
- 31:27that personalized pricing is illegal under all
- 31:29circumstances. You already see it in different
- 31:31markets. My main point is that even as California
- 31:37and many other states consider bills to prohibit
- 31:40this practice, one should not infer from that
- 31:44that it is otherwise legal. There are serious
- 31:48legal risks. You just added another one that's
- 31:50very important. Excellent. That's a great point.
- 31:54So we've talked about using AI and how that can
- 31:56go sideways. But what about AI impersonator fraud
- 32:00or fake review fraud? Can you talk a bit about
- 32:03issues you see with that coming up where a lot
- 32:06of people today really rely on this product's
- 32:09got five stars, it must be great, or this company
- 32:11is wonderful and, in fact, don't realize that
- 32:15maybe there's fake AI reviews behind it that
- 32:18have substance. There's a lot to say about that.
- 32:23So first, big picture. I have serious concerns
- 32:28about AI turbocharging fraud. We've been worried
- 32:31about grandparent scams, where someone calls
- 32:34a grandparent and says, It's Sam. I've been abducted.
- 32:41You need to wire $500 ,000 in crypto immediately,
- 32:46et cetera, et cetera. AI is going to make those
- 32:50types of scams easier to scale and more effective.
- 32:55You could take three seconds or so from the video
- 32:58you're watching now, feed it into one of these
- 33:01voice cloning tools, clone my voice and start
- 33:04calling people. I'm not encouraging anyone to
- 33:06do this. It is illegal calling people and extorting
- 33:09them. And there are many other ways. I already
- 33:15mentioned how people are making false claims
- 33:18about how AI can make people rich. But the whole
- 33:21notion of artificial intelligence is it is artificial
- 33:24intelligence. It's all about impersonating people.
- 33:29So when you have year after year at the FTC impersonation
- 33:33fraud, the number one source of fraud complaints,
- 33:37and then you layer on top, and that's before
- 33:39the rise of AI, and then you layer on top of
- 33:41that a technology perfectly designed to impersonate
- 33:44people at scale, this is a massive, massive problem.
- 33:50You mentioned reviews in particular. AI has benefits
- 33:55here. I certainly have used AI to summarize different
- 34:00reviews people have left. But therein lies the
- 34:03question that you raised. Are the reviews AI
- 34:06is summarizing themselves written by AI tools,
- 34:09or are they actual people? The FTC brought a
- 34:14case end of last year against a company called
- 34:18Ryder, R -Y -T -R, where we alleged the company
- 34:21developed a tool to generate thousands of reviews
- 34:24about a product. And we do not allege that the
- 34:28company committed fraud, but we allege that the
- 34:30company provided the means and instrumentalities
- 34:33for others to commit fraud. And we allege that
- 34:36it was an unfair practice to develop such a tool
- 34:38because it's hard to imagine a legitimate use
- 34:42of a tool to generate a bunch of reviews of a
- 34:45product that AI didn't try on the socks that
- 34:49the human did. Now we can talk about this. That
- 34:56case we brought brought two very strong, not
- 35:00I would say in their persuasiveness, but strong
- 35:03in their tone, dissents from Republican commissioners,
- 35:07one of whom is now chair, happy to talk about
- 35:09that. But we allege that was unfair. And what
- 35:11it speaks to is the problem you're raising. In
- 35:13addition to impersonation fraud, like voice cloning,
- 35:17you can also use AI to generate fake reviews,
- 35:23which is already a huge problem. The FTC tried
- 35:26to curb, we finalized during my tenure a review
- 35:28prohibiting these reviews, but AI can turbocharge
- 35:33them. So let me just talk briefly about the strategy.
- 35:35And then if you're interested, we can talk about
- 35:36the partisan split on that case. So I think it's
- 35:39interesting. First, we wanted to make sure that
- 35:43there were strong penalties for engaging in fraud.
- 35:46So we finalize an impersonation rule that gives
- 35:49the FTC height remedial tools to challenge, it
- 35:52was already deceptive, gives the FTC height remedial
- 35:54tools to challenge impersonation fraud. Second,
- 35:57we finalize a fake review rule, same theory,
- 36:00already deceptive under the FTC Act. This gives
- 36:02the FTC much stronger tools to deter companies
- 36:06from engaging in this practice by allowing the
- 36:09FTC to recover large civil penalties. Third,
- 36:14we launched a voice cloning challenge where we
- 36:18offered a cash award. to folks, we invited people
- 36:22to submit technological and policy solutions
- 36:25to interrupt voice cloning. We got some really
- 36:27exciting submissions. We named the awardees at
- 36:31the end of last year. That generated a lot of
- 36:33enthusiasm, both at the FTC and among academics
- 36:37and technology firms for leveraging American
- 36:40innovation to stop fraud. And the last piece
- 36:43of it, which brings me back to the beginning,
- 36:45is we really pursued an upstream enforcement
- 36:47strategy. We didn't just want to go after the
- 36:51scammers who impersonate grandma or who use voice
- 36:56cloning to impersonate grandma. We wanted to
- 36:59make sure that the companies that were deploying
- 37:00these extremely powerful tools were taking adequate
- 37:04precautions, whether that means know your customer,
- 37:07whether that means verification to the person's
- 37:10voice being cloned is actually that person. We
- 37:14started on that path in enforcement. For reasons
- 37:18I can get into, I am not sure that will continue
- 37:20at the federal level, but it is my hope that
- 37:24states will pick that up. And I'm happy to get
- 37:26into that more if that would be of interest.
- 37:28I think it definitely would. I mean, I think
- 37:30it's important. We've obviously talked about
- 37:32the really important work done under your tenure
- 37:35and previously, but what do you see as sort of
- 37:40taking a crystal ball out and what you've seen
- 37:42since you left in terms of the direction the
- 37:45FTC has gone or potentially may go? Again, not
- 37:50asking you to totally predict the future, but
- 37:53kind of what are you seeing since you've left
- 37:55and changes that are sort of on the horizon?
- 37:57and potentially? Well, I'll start with the good
- 38:00news. And then I'll get to the very disconcerting
- 38:08news. So the good news is A, the CFPB has been
- 38:14destroyed, effectively, or nearly destroyed.
- 38:18I am heartbroken by that. It's terrible for the
- 38:22American economy. It's terrible for American
- 38:24consumers. I love the FTC. I would have spent
- 38:27my whole career there if I could. The FTC can't
- 38:31just, both because of its authority, different
- 38:33authority, different jurisdiction, different
- 38:35resources. The FTC can't just pick up the slack
- 38:38of the CFPB. That's not the focus of this webinar,
- 38:43so I won't get into it, but that's a major problem
- 38:46in the consumer protection landscape. Generally,
- 38:48thank goodness folks California had the foresight
- 38:51to stand up the DFPI, your own mini CFPB in the
- 38:55state. The other states should probably do the
- 38:58same. Back to the FTC. So the FTC has not been
- 39:04defunded or anything like that. But the president
- 39:06illegally fired, in my view, two Democratic commissioners,
- 39:10Commissioner Bedoya, Commissioner Slaughter.
- 39:12They have sued. I expect that case is going to
- 39:15go to the Supreme Court. We can get into that
- 39:19if people are interested. Commissioner Bedoya,
- 39:22who is someone whom I really admire and respect
- 39:25and like as a friend, I was very sad to see just
- 39:29yesterday he announced he was resigning due to
- 39:34his statement could speak for itself. So the
- 39:38F2C as of today is no longer bipartisan. You
- 39:41only have three Republican commissioners actually
- 39:44sitting on the commission. That's a major problem
- 39:46too. Can you explain the way that the FTC is
- 39:53normally set up in terms of the five commissioners,
- 39:55for those who may not know? I was tempted to
- 39:57do that, but I didn't want to go down a rabbit
- 40:00hole, but I'll try to be brief. So there's a
- 40:04myth that it has to be bipartisan. It doesn't
- 40:06have to be bipartisan. What the statute says
- 40:08is no more than three members can be of the president's
- 40:11party, and they can only be removed for cause.
- 40:15For the whole history of the FTC, there have
- 40:17been moments, including during my tenure, when
- 40:19there have only been Democrats or only been one
- 40:22party at the FTC. But the norm has been for the
- 40:26whole 110 -year history of the FTC, there's been
- 40:29a bipartisan commission that's part of an institutional
- 40:32design. The protections for commissioners, the
- 40:37idea that presidents can only fire them for cause,
- 40:40was specifically challenged in the 1930s and
- 40:43the Supreme Court upheld those protections in
- 40:46the famous case, famous to FTC nerds, Humphrey's
- 40:49Executor. The president, Trump, has asserted
- 40:54that Humphrey's Executor is not good law. He
- 40:57is not. in spite of the Supreme Court repeatedly
- 41:00reaffirming it, fired Commissioner Slaughter
- 41:02and Commissioner Bidoya without even an assertion
- 41:04of cause. There are no longer Democrats on the
- 41:08commission. Well, I should say, I believe Commissioner
- 41:13Slaughter was illegally fired and that she therefore
- 41:16still is a commissioner, but in terms of actually
- 41:19in the job voting on matters, she is not. Thank
- 41:24you. Thanks for explaining that. That's really
- 41:26helpful. You talked a bit about rulemaking and
- 41:30some of the things the FTC has done in that space.
- 41:33You've given some examples. Are there other examples
- 41:36you can talk about in terms of rulemaking that's
- 41:38happened or if you anticipate potential changes
- 41:42to rules that are currently in place in the future
- 41:45with the new composition of the commission? Sure.
- 41:49Well, let me actually go back to enforcement,
- 41:51since I didn't quite cover that, and then I'll
- 41:53turn to rulemaking. We have not seen so far in
- 41:58these first six months, feels like longer, of
- 42:04the administration. We have not seen what the
- 42:07privacy enforcement is going to look like, which
- 42:10is interesting to me. A couple data points. Some
- 42:18of the most ambitious privacy cases that the
- 42:20FTC brought during my tenure not only got bipartisan
- 42:23support, they got strong bipartisan support.
- 42:26And I think that's important message for this
- 42:28audience. The GoodRx case, which got a ban on,
- 42:32it didn't say, tell people you're selling, you're
- 42:35sharing their medication information. It prohibited.
- 42:38GoToRacks from sharing medication information
- 42:40for advertising purposes got strong bipartisan
- 42:44support. Commissioner Wilson issued a statement.
- 42:48We haven't talked about this yet, but we brought
- 42:50a series of, I think, really groundbreaking cases
- 42:53against data brokers for mishandling consumers'
- 42:56location data. Those have overwhelmingly been
- 42:59bipartisan, both with Republicans who were on
- 43:02the commission early in the Biden administration
- 43:04and with Republicans still in the commission
- 43:06today. There was some nuance there. There were
- 43:08different counts in different cases that they
- 43:10dissented on. But read their statements. I believe
- 43:14that there is a consensus among commissioners.
- 43:17that data brokers' practice of freely sharing
- 43:21and selling sensitive geolocation data can be
- 43:23unfair under the FTC Act. I believe that every
- 43:26commissioner has said as much. Children's privacy,
- 43:31you see even more bipartisan support. The Epic
- 43:34Games case, we have not talked about yet. The
- 43:36largest children's privacy case we ever brought.
- 43:39Challenged the company's default settings as
- 43:41being unfair under the FTC Act. Recovered more
- 43:44than half a billion dollars. in the monetary
- 43:48relief, got strong bipartisan support. We had
- 43:53the case I mentioned over the summer with the
- 43:55LA District Attorney against this major app,
- 43:58NGL, which banned the company from marketing
- 44:01and services to kids or teens, also got bipartisan
- 44:04support. So you see a lot of bipartisan support
- 44:08for even some of the more far -reaching privacy,
- 44:11especially kids' privacy cases. This year under
- 44:15the new administration, we were planning to hold
- 44:20a workshop on the attention economy and how tech
- 44:23companies were using design techniques to try
- 44:26to keep people hooked, potentially harming kids.
- 44:29The new administration did not abandon that.
- 44:33changed the nature of it quite a bit, made it
- 44:36very focused on big tech and harms from big tech.
- 44:38But they went forward with that workshop. And
- 44:40if you read some of the speeches that the Republican
- 44:42commissioners said, they said that big tech privacy
- 44:46practices and practices around addiction and
- 44:50children were very concerning to them. So I don't
- 44:55exactly know what to make of this new administration.
- 44:58I think They supported many of the most ambitious
- 45:03privacy cases over the last four years. I think
- 45:07they'll continue to be aggressive on kids, including
- 45:11on our new children's privacy rule, which was
- 45:13unanimously supported. That was finalized beginning
- 45:17of this year. I don't know where they will be
- 45:20on data brokers or some of these other issues.
- 45:23We just have not seen any enforcement yet. And
- 45:27by the way, that itself is concerning. The fact
- 45:29that we are coming on six months into this administration,
- 45:35and I could be mistaken, but I don't think they
- 45:37have filed a single privacy case. I personally
- 45:42don't like the signal that that sends. I would
- 45:46not assume that they're not going to bring in
- 45:48privacy cases in the future. As you know, Christy,
- 45:50there's always a ramp up with the new administration
- 45:52while they're vetting the investigations already
- 45:54underway and deciding whether it comports with
- 45:57their own approach to the law. Right. Right.
- 46:00No, that's really helpful. Oh, rulemaking. Should
- 46:04I? Go back to that. Yeah, they're not going to
- 46:07do. So we. I won't get into the long FTC history
- 46:11of rulemaking, though. I think it's interesting.
- 46:12On privacy rulemaking in particular, the big
- 46:15thing we did not do that I wish we could have
- 46:17done, that I think we would have done if we had
- 46:20had more time, was a comprehensive privacy rule
- 46:23under the FTC Act. We solicited comments. I can't
- 46:28get into non -public information about work that
- 46:30was done or that was not done, but it is a real
- 46:33regret of mine that we were not able to issue
- 46:35a rule. I think a rule, baseline privacy protections
- 46:39are long overdue. Because many states don't have
- 46:46any privacy laws, let alone privacy laws as strong
- 46:48as California's. I think the chance of this administration
- 46:51moving forward with the privacy rule given, I
- 46:54mean, you have an executive order that says for
- 46:57every rule promulgated, you need to repeal 10.
- 47:02You don't have 10 rules. I mean, you barely have
- 47:0510 rules at the FTC, as you know, to repeal.
- 47:08So I don't think there's any chance this administration
- 47:11moves forward with a new privacy rule. But there
- 47:15was privacy rulemaking that we did that. I think
- 47:18will survive. Graham Leach -Bliley, we significantly
- 47:22updated the safeguards rule, which has to do
- 47:25with data security standards for non -bank financial
- 47:28institutions. We significantly updated Children's
- 47:31Online Privacy Protection Act, got unanimous
- 47:33support for changes we made, especially around
- 47:37making it easier for parents to opt out of targeted
- 47:39advertising, prohibiting indefinite retention,
- 47:42and more. And we updated significantly our health
- 47:46breach notification rule, which we haven't talked
- 47:49about yet, but I think it's really important
- 47:51for companies that are operating apps or other
- 47:54services that collect sensitive health information.
- 47:57If it's being shared without consumers authorization,
- 48:00that can be a breach under the health breach
- 48:02notification, even if it wasn't stolen, even
- 48:04if it was just shared without. consumer authorization.
- 48:08That can be considered a breach under the health
- 48:10breach certification rule, and companies can
- 48:12be sued and face civil penalties for not reporting
- 48:15that breach to consumers and to the FTC. That's
- 48:18another rule that we updated that so far has
- 48:21not been touched. Plus, as I mentioned, the impersonation
- 48:25rule, the fake review rule, which are going to
- 48:27be critical tools to combat fraud. but the administration
- 48:33has already been enforcing both rules. The junk
- 48:35fee rule, which I talked about surveillance pricing
- 48:38earlier, the junk fee rule requires all -in pricing.
- 48:41The administration has already indicated they're
- 48:43going to be enforcing that rule. I think with
- 48:45the rules we did, we're seeing continuity, which
- 48:47is really great to see. But you're not going
- 48:50to see, I don't think any new rules promulgated
- 48:52under this administration, at least on privacy.
- 48:56Right, right. No, that's helpful. You talked
- 48:58a little bit about the healthcare privacy rules,
- 49:02and I know there have been a number of cases
- 49:03the FTC has brought in the healthcare space around
- 49:06privacy. Can you talk a bit about that and what
- 49:09people should know and understand if they're
- 49:11working in the healthcare space and dealing with
- 49:14private information and what they might want
- 49:16to think about in that context? So when the Dobbs
- 49:19decision came down in 2022, I think that was
- 49:23a wake -up call for a lot of Americans about
- 49:27privacy not being an abstraction, that, oh, you
- 49:31mean the government can actually go to a data
- 49:34broker and purchase information on all my whereabouts
- 49:38all the time, what church I'm going to, what
- 49:40doctor I'm going to, what school I'm going to,
- 49:44where I'm spending my days, when I'm crossing
- 49:47state lines. Not new to the FTC as you know Christina
- 49:53this is something we had been sound the alarm
- 49:55about for years. that the, especially the surreptitious
- 49:58collection of sense. You know, when people download,
- 50:00I know all the lawyers on this crowd know this,
- 50:02when you download an app and it's free, you know,
- 50:06you're the product, right? Your location is being,
- 50:09or your information is likely being harvested
- 50:12by that app and sold. So there's this vast underground
- 50:16market for people's sensitive geolocation data
- 50:19that can reveal information since you asked about,
- 50:22it can reveal all sorts of things. But among
- 50:25other things, it can reveal, information about
- 50:27people's healthcare, what healthcare they're
- 50:29seeking. Similarly, and the FTC, as I alluded,
- 50:34you took a very aggressive approach to that.
- 50:36We sued data broker after data broker, alleging
- 50:39that they needed to start filtering out sensitive
- 50:43data, data that reveals, for example, people's
- 50:46visits to reproductive health centers, and not
- 50:49straight out prohibition on onward sale or sharing
- 50:52of that data. So that was one way we tried to
- 50:54protect health privacy. Another way was we went
- 50:57after health data directly. So I had been talking
- 50:59about location data that can reveal people's
- 51:01health. We also brought cases like pre -mom,
- 51:04cerebral. good Rx, better help. These were all
- 51:08cases where companies collected people's sensitive
- 51:10health information, what mental health treatments
- 51:13they were seeking, what prescriptions they were
- 51:16taking, whether they were looking to have a child,
- 51:21and sharing that information with different social
- 51:25media companies so that advertisers could target
- 51:29their users. We alleged in some cases that this
- 51:33violated the health breach notification rule
- 51:35I mentioned earlier. We alleged in all cases
- 51:38that these practices were unfair under the FTC
- 51:42Act. And my strong advice to companies that collect
- 51:47sensitive health information, whether you agree
- 51:51with me that it's unfair or not, If consumers
- 51:54find out about it, and we know that this happened,
- 51:57they're going to be really upset. I really believe
- 52:00it is a breach of trust. If you really want to
- 52:04share some insensitive health information, you
- 52:06should be at a minimum getting affirmative express
- 52:09consent. And that means if consumers... That
- 52:13means consumers can opt in. And if they don't
- 52:14want to share it, they shouldn't have to share
- 52:16it. I think this is really visceral for people.
- 52:18I don't think it's a partisan issue. I don't
- 52:21think people want to have what medication they're
- 52:24taking. They don't want that information being
- 52:27sold. So we took a multifaceted approach to health
- 52:30privacy, rulemaking, enforcement guidance. But
- 52:34there's still a lot of work to do. And I certainly
- 52:36hope this administration picks it up. Yeah, no,
- 52:42that's really helpful. So we do have a couple
- 52:48of questions that have come in. So maybe we'll
- 52:50ask those before we run out of time. But we have
- 52:54a question about, we talked about surveillance
- 52:56pricing where someone might have a higher price.
- 52:59Someone asked about surveillance discounts. Is
- 53:02that the same analysis a company should do if
- 53:04they decide? that Sam likes to hover over this
- 53:10page 15 times and so maybe we're going to give
- 53:12him a discount so he decides to put it in his
- 53:15basket. There's a question about whether that
- 53:17surveillance discounting may have some similar
- 53:19implications or things to be concerned about.
- 53:22I think it depends. Questions I would ask include,
- 53:27is this a discount available to all consumers?
- 53:30Is this a fast food chain saying, discount on
- 53:34fries from 11 to 12 o 'clock, or is this a discount
- 53:38to individual consumers based on data that that
- 53:42app was collecting about them? I would ask, is
- 53:45it an actual discount? If I'm giving you, Christina,
- 53:50a 20 % discount, and I'm giving everyone else
- 53:53a discounted price, 30%, 10%, 40%, but no one's
- 53:57actually paying the... the reference price, the
- 54:01real price. It's a long standing FTC law and
- 54:05I think law in just about every state that that
- 54:07is a deceptive practice. So I think it really
- 54:10depends on the circumstance. I saw that the same
- 54:14person asked about Uber surge pricing. I don't
- 54:20want to get into it. how Uber's algorithm works.
- 54:23That would be a question for Uber. The question
- 54:27was, is Uber surge pricing not personalized pricing?
- 54:29I mean, it can be personalized pricing. And again,
- 54:34if you're a company and you want to start engaging
- 54:39in this type of practice, go out and ask your
- 54:42customers whether they like Uber surge pricing.
- 54:45Go out and ask your customers whether they like
- 54:48booking a flight. at the last minute. Setting
- 54:54aside the legal risk here, I think it's a real
- 54:58source of consumer frustration by me as well.
- 55:02Absolutely. Can you talk a little bit about the
- 55:06FTC and partnerships? I know you talked about
- 55:08working with the DA, but maybe for the audience
- 55:12to have an understanding about circumstances
- 55:15and. instances where the FTC may choose to partner
- 55:19with a state or a local agency and how that might
- 55:23benefit consumers in terms of that kind of a
- 55:25practice. So I'm an alum of an AG's office. I
- 55:29love state work. I love municipal work too. States
- 55:33are on the front lines. They have the best sense
- 55:35of what the challenges are consumers are facing
- 55:37in their jurisdictions. One of my priorities
- 55:40as director was strengthening partnerships with
- 55:46states. We issued a report on how we were trying
- 55:48to do that. Our regional offices, including one
- 55:51in LA and one in San Francisco, are really on
- 55:54the front lines of that. Christina is an alum.
- 55:57But as Christina also knows, state federal collaboration
- 56:00certainly did not begin with my tenure. It's
- 56:02actually how the two of us met was some years
- 56:04ago on a case Christina was working on at the
- 56:07FTC and I had some knowledge of in state government.
- 56:11FTC partners with states all the time. That became
- 56:15more urgent in 2021 when the Supreme Court handed
- 56:18down a decision called AMG that basically stopped
- 56:22the FTC from getting refunds for people in federal
- 56:26court unless a rule was violated. In other words,
- 56:30if you engage in unfair deceptive practices,
- 56:34the California DOJ could go into court and get
- 56:36money back to victims. FTC cannot go into federal
- 56:39court and get money back to victims unless a
- 56:42specific rule had been violated. So because of
- 56:44that ruling, the need and the importance of partnering
- 56:48with states only grew. And one of the things
- 56:50we started doing in my tenure was partnering.
- 56:53We had always partnered with states on fraud
- 56:55and other areas. We really started partnering
- 56:58with states on more tech cases. including a mention
- 57:02with the LA District Attorney. We brought CRI
- 57:06genetics with California. We really tried to
- 57:12strengthen our relationship both with state AGs
- 57:14and with state privacy enforcers to make sure,
- 57:19A, that we were maximizing the remedies we could
- 57:21recover in court. make sure we could make consumers
- 57:25whole. But B, I think we really benefited from
- 57:29hearing from states about what the biggest issues
- 57:31their consumers were facing and how could we
- 57:33as federal enforcers support their work. Fantastic.
- 57:37Yeah, I mean, I think having also come from a
- 57:39state agency's office before, I think those partnerships
- 57:41are really important and being able to address
- 57:44some of the AMG issues, but also being able to
- 57:46have those partnerships where state and local
- 57:48agencies can be part of it and really work to
- 57:51help protect their consumers as well, using their
- 57:53consumer protection laws, which often have penalties
- 57:57and pretty significant restitution pieces too,
- 58:01I think are really another important tool. And
- 58:03I think it's great that the FTCs look to pivot.
- 58:06And I think that's one of the messages that you've
- 58:08had here today is sort of the importance of the
- 58:11FTC looking at the landscape and pivoting, whether
- 58:13it's through rulemaking or whether this you're
- 58:15using different parts of section five of the
- 58:17FTC Act or partnerships as a way to just meet
- 58:20the ultimate goal of sort of helping consumers
- 58:23be protected and be in a fair marketplace, right?
- 58:27Where they can get sort of the best results.
- 58:29So I think that's really important and hats off
- 58:33to you for all of your great work. while you
- 58:34were there, Sam. I mean, really important cases
- 58:37came out during your tenure. We're getting close
- 58:40to the end, so I'll put it back to you to see
- 58:43if you have any closing thoughts or anything
- 58:45you want to add to sort of the audience to consider
- 58:48in this privacy AI space as we get to the last
- 58:51few minutes. So I don't need to tell Californians,
- 58:55especially those of you in the LA area, that
- 58:59the November election led to a very dramatic
- 59:04change in power in Washington, DC. But for those
- 59:09of you who practice in the privacy AI space,
- 59:14I would not assume, I strongly would urge you
- 59:16not to assume that we're just going to go back.
- 59:18to 2016 or 2020 in terms of how the FTC views
- 59:24its mandate on privacy and AI. There were strong
- 59:29disagreements among the FTC commissioners about
- 59:32certain cases, but the Overton window has shifted.
- 59:36I really believe that very fundamentally. You
- 59:40now have bipartisan consensus with strong support
- 59:43in Congress that data brokers are mishandling
- 59:46our sensitive geolocation data, and that can
- 59:50be unfair to the FTC Act. But health apps...
- 59:54need to use far more caution in handling their
- 59:57sensitive health data. And that can be actionable
- 1:00:00under health certification rule or be unfair
- 1:00:03under the FTC Act. You have a sea change in kids
- 1:00:08protections, where this growing state, not only
- 1:00:13growing state interest in enforcing children's
- 1:00:17privacy laws, but strong bipartisan consensus
- 1:00:20at the FTC, both for expanding and strengthening
- 1:00:24the FTC's children's privacy rule and enforcing
- 1:00:27that rule aggressively. And you've heard that
- 1:00:29rhetoric very consistently. So in spite of the
- 1:00:32dramatic change we've seen in Washington, generally,
- 1:00:36I think when it comes to basic FTC compliance
- 1:00:39issues, you are not going to see a 180. You'll
- 1:00:42see changes, to be sure, but we're not just going
- 1:00:45to go back to the notice and choice era of the
- 1:00:48past. We are in a new world, I think, where companies
- 1:00:51need to be thinking about not just whether they
- 1:00:53disclose what they're doing with people's data
- 1:00:56or how they're using AI tools. They need to be
- 1:00:58thinking about are our collection practices,
- 1:01:02are our sharing practices, are our tools harming
- 1:01:05people? And if they are, that should raise concerns
- 1:01:08whether we're engaging in an unfair practice
- 1:01:10under the FTC Act. That would be my, I think,
- 1:01:13most important message for people trying to stay
- 1:01:15on the right side of the law, which I know as
- 1:01:19all of you. Absolutely, absolutely. Well, thank
- 1:01:22you so much for your time, Sam. This was really
- 1:01:25enlightening and we really appreciate your contributions
- 1:01:29to the Beverly Hills Bar. With that, I think
- 1:01:31we'll close out the session. Thank you so much,
- 1:01:33everyone, for joining. Thank you for being here.
- 1:01:35Thank you for joining and thanks for having me.
- 1:01:37It's been great to be here.