WEBVTT - FTC Chair Lina Khan: The Better Offline Interview 0:00:02.360 --> 0:00:03.520 Cool Media. 0:00:05.160 --> 0:00:08.360 Hello, and welcome's a Better offline. I'm your host ed zietron. 0:00:16.640 --> 0:00:16.880 As. 0:00:19.760 --> 0:00:21.799 I am joined by the Chair of the Federal Trade Commission, 0:00:21.840 --> 0:00:23.919 Lena Cohn. Lena, thank you so much for joining me. 0:00:24.600 --> 0:00:25.319 Great to be here. 0:00:25.960 --> 0:00:29.480 So on a simple level, what does the FTC actually do? 0:00:30.120 --> 0:00:34.720 So, the FTC enforces America's anti trust and consumer protection laws, 0:00:35.000 --> 0:00:37.000 and so these are the laws that try to ensure 0:00:37.040 --> 0:00:40.839 that our markets are fair and honest and competitive, and 0:00:40.880 --> 0:00:44.479 so that means we're looking to protect people from illegal monopolies. 0:00:44.800 --> 0:00:47.360 We're trying to protect people from all sorts of frauds 0:00:47.440 --> 0:00:50.640 and scams. We're trying to protect people's privacy and wanting 0:00:50.680 --> 0:00:52.720 to make sure that they're not subjected to all sorts 0:00:52.720 --> 0:00:56.280 of unfair practices. We were created one hundred and ten 0:00:56.360 --> 0:01:00.880 years ago, and so it's a pretty big mandate. We 0:01:00.920 --> 0:01:04.720 have oversight across markets across the entire economy for the 0:01:04.720 --> 0:01:08.039 most part, with some exceptions like in areas such as 0:01:08.080 --> 0:01:11.840 banks and airlines. So it's a really big job that 0:01:11.880 --> 0:01:14.640 this agency does and it's been thrilling to get to 0:01:14.640 --> 0:01:14.920 do it. 0:01:15.760 --> 0:01:18.960 And it feels like you're a lot more aggressive than 0:01:19.160 --> 0:01:22.240 perhaps anti monopoly and anti consolidation has been in the past, 0:01:22.280 --> 0:01:23.000 what led to them. 0:01:24.040 --> 0:01:26.720 So I came into this job wanting to make sure 0:01:26.840 --> 0:01:30.640 that we were a fully using the tools that Congress 0:01:30.680 --> 0:01:33.360 gave us. As I noted, this agency has been around 0:01:33.440 --> 0:01:37.080 for o first century. Our agency was given some pretty 0:01:37.120 --> 0:01:41.319 important authorities and tools, but over the decades there had 0:01:41.360 --> 0:01:44.880 been a drift away from fully using all of the 0:01:44.880 --> 0:01:48.920 tools that we have, from fully embracing the law on 0:01:49.040 --> 0:01:51.400 the books. And so the first thing we did was 0:01:51.480 --> 0:01:53.840 kind of come in and just make sure that we 0:01:53.880 --> 0:01:56.600 are mapping out what are all the tools we actually 0:01:56.600 --> 0:01:59.320 have and how do we make sure we're faithfully using those. 0:02:00.480 --> 0:02:03.280 The second thing that's been important is wanting to make 0:02:03.320 --> 0:02:07.120 sure that we're enforcing the law and taking action in 0:02:07.200 --> 0:02:12.200 response to how markets actually work today, and that meant 0:02:12.400 --> 0:02:16.600 making sure we were having top expertise Internally. We started 0:02:16.600 --> 0:02:20.680 a new Bureau of Technologists that you know, brought data scientists, 0:02:20.760 --> 0:02:24.119 data engineers, people who fully understand the ins and outs 0:02:24.160 --> 0:02:29.240 of how algorithms work. Previously, companies would view lawsuits as 0:02:29.360 --> 0:02:32.640 just a cost of doing business, and especially in digital 0:02:32.680 --> 0:02:35.760 markets where there's such a focus on chasing scale and 0:02:35.840 --> 0:02:38.880 chasing growth, you know, there was a bit of a 0:02:38.880 --> 0:02:43.240 ask for forgiveness later mentality, rather than everybody abiding by 0:02:43.240 --> 0:02:45.880 the law and making sure Americans were getting their rights 0:02:45.919 --> 0:02:46.560 on the front end. 0:02:47.440 --> 0:02:50.400 So how do you curtail that instinct where they have 0:02:50.520 --> 0:02:53.160 so much money that they can just kind of absorb 0:02:53.240 --> 0:02:54.840 the fee, sorry, the fine. 0:02:54.880 --> 0:02:58.840 I mean, yeah, it's a good question, and it's especially 0:02:58.880 --> 0:03:02.680 important because a few months before I came into this role, 0:03:03.280 --> 0:03:07.240 the courts had actually further limited the FTC's ability to 0:03:07.280 --> 0:03:11.240 get money. We mitigated that in a couple of ways. First, 0:03:11.320 --> 0:03:13.600 we used the tools we did have to their full 0:03:13.639 --> 0:03:18.079 extent to still get record breaking sums. We also made 0:03:18.120 --> 0:03:22.240 sure we were looking at things like individual accountability, and 0:03:22.320 --> 0:03:27.800 so if a specific individual executive was involved in directing 0:03:28.639 --> 0:03:32.000 or participating in the law breaking, the lawsuit can be 0:03:32.040 --> 0:03:35.080 brought not just against the company but also against the 0:03:35.120 --> 0:03:39.600 individual executive. And so we've done that in several instances 0:03:40.280 --> 0:03:43.600 where we had evidence that high level execs knew what 0:03:43.680 --> 0:03:46.440 was going on, had the ability to stop it, and didn't. 0:03:46.960 --> 0:03:49.480 And there's been a lot of empirical work that shows 0:03:49.480 --> 0:03:55.800 that actually holding individual executives accountable is really important for 0:03:56.040 --> 0:03:59.800 deterring law breaking in the first instance. So we've done that. 0:04:00.080 --> 0:04:04.200 We've also made sure that when companies, for example, have 0:04:04.360 --> 0:04:09.440 used illegal practices to hoard people's data, that they're having 0:04:09.480 --> 0:04:12.120 to actually get rid of that data or get rid 0:04:12.200 --> 0:04:14.880 of the models that we're trained on the data. So 0:04:14.920 --> 0:04:18.560 we're actually getting remedies that make sense for today's markets. 0:04:19.440 --> 0:04:22.159 So you file the case against Amazon? What is the case? 0:04:22.200 --> 0:04:23.680 What's going on over there? Because I know you have 0:04:23.720 --> 0:04:26.440 some history when at Yale Law School you wrote that 0:04:26.720 --> 0:04:29.840 the antitrust paradox. I believe it, like, what's the case 0:04:29.839 --> 0:04:30.560 against Amazon? 0:04:32.440 --> 0:04:35.839 So just to step back, what we see with digital 0:04:35.880 --> 0:04:39.760 platforms in particular is that there can be a life cycle. 0:04:40.320 --> 0:04:44.000 Right initially, when a digital platform is entering the market, 0:04:44.400 --> 0:04:48.840 they're really focused on chasing scale and building up a 0:04:48.960 --> 0:04:51.680 large customer base on one side of the market, so 0:04:51.760 --> 0:04:54.159 then they can also attract a big customer base on 0:04:54.200 --> 0:04:57.280 the other end, and they want to basically achieve, you know, 0:04:57.360 --> 0:05:01.719 the flywheel of accelerated growth in them. Mentum that, especially 0:05:01.760 --> 0:05:04.719 in digital markets, once you've kind of hit that sweet 0:05:04.720 --> 0:05:09.800 spot of sufficient scale, there can be enormous acceleration. But 0:05:09.880 --> 0:05:14.440 what we've seen is that once firms achieve that accelerated growth, 0:05:14.680 --> 0:05:18.440 and if they're successful at knocking out their arrivals, we 0:05:18.480 --> 0:05:22.480 can see a different phase kick in where the platform 0:05:22.600 --> 0:05:26.440 starts looking to milk the monopoly right, and we can 0:05:26.480 --> 0:05:29.640 see that in a few ways, the company can start 0:05:29.839 --> 0:05:34.839 charging higher prices. Sometimes they can make service worse. Corey 0:05:34.880 --> 0:05:37.920 doctor Rowe has coined this term and shoitification around just 0:05:38.040 --> 0:05:41.479 the degradation of the services that you see. And so 0:05:41.600 --> 0:05:45.200 the lawsuit that we've brought is about Amazon as it 0:05:45.279 --> 0:05:50.440 is in this second phase, where basically we alleged they 0:05:50.480 --> 0:05:55.640 have height prices for both buyers on Amazon's platform as 0:05:55.680 --> 0:05:59.320 well as the small businesses that sell through its platform, 0:05:59.440 --> 0:06:02.200 where Amazon now collects one out of every two dollars. 0:06:02.800 --> 0:06:06.240 We also note that Amazon has kind of cluttered its 0:06:06.320 --> 0:06:10.960 search results page with irrelevant ads as a very deliberate 0:06:10.960 --> 0:06:14.080 strategy to be able to basically milk more ad revenue 0:06:14.120 --> 0:06:17.760 from the platform even when it's not serving shoppers and 0:06:17.839 --> 0:06:22.280 serving consumers. And so we allege that Amazon's only been 0:06:22.320 --> 0:06:25.960 able to make things worse in this way for people 0:06:26.120 --> 0:06:30.400 because they've architected this anti competitive scheme to block out 0:06:30.520 --> 0:06:34.560 rivals through illegal tactics. And so that's what the lawsuit 0:06:34.680 --> 0:06:36.920 is about. It's going to go to trial in a 0:06:36.960 --> 0:06:39.760 couple of years, and we brought it with a whole 0:06:39.760 --> 0:06:41.520 bunch of state attorneys general as well. 0:06:42.200 --> 0:06:46.040 Lovely So some one of the core thrusts you made 0:06:46.040 --> 0:06:49.960 in your paper. Amazon's antitrust paradox is that they prioritized 0:06:50.720 --> 0:06:54.359 growth over being a profitable company. This feels like a 0:06:54.440 --> 0:06:57.520 problem across the tech firms, all of them, they all 0:06:57.600 --> 0:07:00.960 seem to be growth hungry. I feel do you feel 0:07:00.960 --> 0:07:03.400 that this is anti competitive at its core because growth 0:07:03.400 --> 0:07:06.080 doesn't seem to be about better, just seems like more. 0:07:07.640 --> 0:07:10.800 Yeah, it's a really good question. And the antitrust laws 0:07:10.880 --> 0:07:15.280 in no way say that chasing scale or chasing growth 0:07:15.640 --> 0:07:19.360 is illegal, or even that being big is illegal. It 0:07:19.480 --> 0:07:23.640 really comes down to what are the mechanisms of chasing 0:07:23.680 --> 0:07:27.320 that growth or the mechanisms of becoming big? Right, and 0:07:27.360 --> 0:07:30.720 so in the same ways that we would recognize, hey, 0:07:31.080 --> 0:07:33.480 like going across You know, if you have a pizza 0:07:33.480 --> 0:07:35.760 shop that goes across the street and sets fire to 0:07:35.840 --> 0:07:38.800 its rival pizza shop, you know that's not fair competition. 0:07:39.360 --> 0:07:41.520 That's an extreme example, but there are all sorts of 0:07:41.560 --> 0:07:45.400 other mechanisms that the law say are not fair means 0:07:45.400 --> 0:07:48.760 of competing, and so that's what the law does. It 0:07:48.800 --> 0:07:53.120 tries to distinguish what is fair play versus not when 0:07:53.160 --> 0:07:56.320 you're looking to, you know, compete and grow and get big. 0:07:57.800 --> 0:08:01.560 And so chasing scale and growth is not in any 0:08:01.600 --> 0:08:04.320 way by itself illegal and instead is a really kind 0:08:04.320 --> 0:08:08.640 of core part of what incentivizes business. Is What we 0:08:08.800 --> 0:08:13.600 have seen though, is if law enforcers are not effective 0:08:13.640 --> 0:08:19.160 and are not timely, chasing growth can happen through playing 0:08:19.160 --> 0:08:23.560 fast and loose with the rules, especially if companies know that, Okay, 0:08:23.560 --> 0:08:26.080 sure I might get hit with a lawsuit in five years, 0:08:26.120 --> 0:08:28.840 but guess what, I will have hit my targets by then, 0:08:29.080 --> 0:08:32.160 I'll be the biggest firm. I'll have achieved dominance, and 0:08:32.240 --> 0:08:35.839 so I can worry about that later. From a business perspective, 0:08:35.880 --> 0:08:39.160 it's actually better for the company to play fast and 0:08:39.160 --> 0:08:42.120 loose with the rules so that they gain that dominance 0:08:42.120 --> 0:08:44.640 and then deal with it later. That's really problematic as 0:08:44.679 --> 0:08:47.080 a mindset for law enforcement, and so that's something that 0:08:47.120 --> 0:08:49.160 we've been trying to figure out. How do we devise 0:08:49.200 --> 0:08:52.680 strategies to be much more assertive, be much more forward leaning, 0:08:52.840 --> 0:08:55.720 be much more timely, So that we're not having to 0:08:55.760 --> 0:08:58.679 do that clean up five years, ten years after. 0:08:58.440 --> 0:09:02.079 The fact, How does the FTC have to actually tackle 0:09:02.200 --> 0:09:06.320 anti trust violators? And indeed within those shortened time scales. 0:09:06.559 --> 0:09:13.520 So we oversee various facets of firms behavior. One is 0:09:13.600 --> 0:09:17.000 if they are looking to merge or pursue an acquisition, 0:09:17.600 --> 0:09:20.400 if it's above you know, one hundred and nineteen million dollars, 0:09:20.440 --> 0:09:22.760 it gets reported both to the FTC and to the 0:09:22.800 --> 0:09:25.280 Anti Trust Division, and we can figure out is this 0:09:25.320 --> 0:09:29.079 an anti competitive transaction? Is this basically going to reduce competition? 0:09:30.200 --> 0:09:32.960 And then we can also look at companies business tactics, 0:09:33.080 --> 0:09:36.800 especially you know, if they've become monopolies or are kind 0:09:36.840 --> 0:09:40.520 of using their dominance in illegal ways, and so we 0:09:40.520 --> 0:09:44.600 can bring lawsuits. We can also issue rules, and we've 0:09:44.640 --> 0:09:47.840 been very active in issuing rules, including in areas like 0:09:48.559 --> 0:09:54.560 fake reviews, you know, taking on things like government impostors. 0:09:55.360 --> 0:09:58.200 We just finalized a rule taking on junk fees in 0:09:58.240 --> 0:10:02.240 the economy. These you know, hidden fees that companies just 0:10:02.320 --> 0:10:04.920 show at the very end that you don't know what 0:10:04.960 --> 0:10:08.440 you're exactly paying for, but suddenly the price of the 0:10:08.480 --> 0:10:11.520 ticket is almost doubled and you have no choice, and 0:10:11.600 --> 0:10:14.280 so we've issued a series of rules, we've brought a 0:10:14.320 --> 0:10:18.040 series of lawsuits. We can also do various studies of 0:10:18.160 --> 0:10:21.360 the market, including in areas where things are moving very quickly. 0:10:21.920 --> 0:10:25.160 So we a few months ago launched a study into 0:10:25.160 --> 0:10:29.040 surveillance pricing, this idea that firms could be charging every 0:10:29.080 --> 0:10:33.160 single individual a standalone price based on the data they 0:10:33.200 --> 0:10:35.200 have on you. So these are just some of the 0:10:35.200 --> 0:10:36.560 tools and authorities we have. 0:10:47.400 --> 0:10:50.640 So let's still call official intelligence. From a regulatory perspective, 0:10:51.040 --> 0:10:52.439 what are your major concerns. 0:10:53.679 --> 0:10:56.040 So the biggest concern for me, just right at the 0:10:56.040 --> 0:10:59.520 beginning was making sure we were making very very clear 0:11:00.160 --> 0:11:03.800 to the market that there was no AI exemption from 0:11:03.800 --> 0:11:06.280 the laws on the books. And this can sound like 0:11:06.320 --> 0:11:09.440 a basic point, but it seemed incredibly important because we 0:11:09.559 --> 0:11:13.760 have seen, especially when new technologies come onto the market, 0:11:14.120 --> 0:11:17.000 we can see an effort by firms to say, Okay, 0:11:17.000 --> 0:11:19.679 sure you have these age old laws on the books, 0:11:20.160 --> 0:11:24.079 but this new technology is different, and that means these 0:11:24.160 --> 0:11:27.600 laws don't apply or they have to apply in dramatically 0:11:27.640 --> 0:11:31.120 different ways. And it's really important for enforcers to hear 0:11:31.160 --> 0:11:33.880 out those arguments, you know, make sure we understand how 0:11:33.880 --> 0:11:37.520 the technologies are working, but not get so dazzled by 0:11:37.559 --> 0:11:40.559 the novelty to say, okay, well, I guess we can 0:11:40.600 --> 0:11:46.120 totally abandon our core laws and suddenly we're saying, okay, 0:11:46.720 --> 0:11:49.920 fraud is illegal, but if you do fraud with AI 0:11:50.200 --> 0:11:53.720 sometimes somehow that's okay. Right, That's not a position we 0:11:53.760 --> 0:11:54.320 want to be in. 0:11:54.480 --> 0:11:57.840 Companies are making like that kind of argument. They were 0:11:57.840 --> 0:12:01.000 attempted to sites that, not that specifically, but they trying 0:12:01.040 --> 0:12:01.400 to dodge. 0:12:03.200 --> 0:12:05.760 Yeah, we do see arguments that, you know, the traditional 0:12:05.880 --> 0:12:11.280 laws around unfair practices or deceptive practices applied differently. The 0:12:11.360 --> 0:12:14.200 other big thing is, you know, America has been so 0:12:14.320 --> 0:12:20.239 fortunate to be such a key home of breakthrough innovations historically, 0:12:20.840 --> 0:12:23.640 and to my mind, a key ingredient of that has 0:12:23.720 --> 0:12:28.400 been making sure we have fair competitive markets where you 0:12:28.480 --> 0:12:32.160 actually do have somebody with a great new idea, an 0:12:32.280 --> 0:12:35.120 upstart that's able to come in, that's able to fairly 0:12:35.200 --> 0:12:39.440 compete and knock out the incumbents. That's how historically we 0:12:39.520 --> 0:12:44.000 have gotten real major technological progress in our country. And 0:12:44.280 --> 0:12:46.280 what I wanted to make sure we were being very 0:12:46.280 --> 0:12:50.679 clear eyed about is the risk of a few large 0:12:50.760 --> 0:12:54.960 dominant incumbents locking in the market in a way that 0:12:55.160 --> 0:12:58.080 shuts out a lot of that innovation and a lot 0:12:58.120 --> 0:13:03.080 of that next generation breakthrough that historically we've seen come 0:13:03.280 --> 0:13:07.480 purely from fair competition and open markets. So it was 0:13:07.520 --> 0:13:10.680 about wanting to make sure that companies are not using 0:13:10.679 --> 0:13:15.000 their existing heft to immediately lock up the market and 0:13:15.640 --> 0:13:18.560 shut out the innovation that we should be able to see. 0:13:19.960 --> 0:13:22.719 So, with that in mind, are you worried about the 0:13:22.760 --> 0:13:25.800 consolidation we're seeing right now in the generative AI market? 0:13:25.840 --> 0:13:29.520 Because to build these models, the foundation models like Claude 0:13:29.559 --> 0:13:32.400 three or GPT or LAMA or what have you, it 0:13:32.440 --> 0:13:36.600 requires financial heft. How do you avoid the monopoly that 0:13:36.679 --> 0:13:38.760 naturally comes out of the fact that you need billions 0:13:38.760 --> 0:13:40.920 and billions of dollars to even make one of these. 0:13:42.520 --> 0:13:45.680 Yeah, it's a key question, and one approach that we've 0:13:45.720 --> 0:13:50.560 taken is making sure that we are actually going layer 0:13:50.679 --> 0:13:53.880 by layer in the market. Right, And so you want 0:13:53.880 --> 0:13:58.080 to understand, Okay, what's happening with chips, what's happening with 0:13:58.240 --> 0:14:02.320 cloud and compute, what's happening with the models. You want 0:14:02.320 --> 0:14:06.320 to understand what are the key economic properties of each 0:14:06.720 --> 0:14:10.720 point in the stack and then understand, Okay, do we 0:14:10.800 --> 0:14:16.000 see this consolidation purely because they're such high fixed costs 0:14:16.160 --> 0:14:20.240 and you know, that's just going to be the state 0:14:20.280 --> 0:14:23.200 of the market. And how do we make sure that 0:14:23.280 --> 0:14:26.320 the degree you do have dominant players in some of 0:14:26.320 --> 0:14:30.440 these layers, they're not using that dominance to then squelch 0:14:30.560 --> 0:14:33.760 out competition in a layer where you should actually be 0:14:33.840 --> 0:14:38.080 able to see more innovation, more competition. And so we're 0:14:38.120 --> 0:14:42.600 looking at, for example, some of the partnerships and investments 0:14:43.000 --> 0:14:49.160 between the big cloud providers and some of the newer firms. Famously, 0:14:49.240 --> 0:14:53.000 you know, the Microsoft Open Ai partnership. Amazon and Google 0:14:53.080 --> 0:14:56.160 have also made some investments, and we want to understand, 0:14:56.160 --> 0:15:00.280 you know, are these truly kind of independent companies. Are 0:15:00.280 --> 0:15:03.320 these able to be independent firms in terms of the 0:15:03.360 --> 0:15:07.520 competitive decision making some of the key strategic calls when 0:15:07.640 --> 0:15:11.200 so much investment is being made. What is the structure 0:15:11.240 --> 0:15:14.040 of the contracts, What are some of the key requirements? 0:15:15.000 --> 0:15:18.120 Are there kind of interlocking director it's here in ways 0:15:18.160 --> 0:15:22.960 that create conflicts that are problematic? Are there contractual you know, 0:15:23.120 --> 0:15:26.600 tying provisions here? So that's just the universe of what 0:15:26.640 --> 0:15:29.080 we're looking at, But of course, it's a fast moving 0:15:29.720 --> 0:15:32.920 market and we're very fortunate to have just top tier 0:15:33.160 --> 0:15:36.800 technological talent in house to be letting us push forward here. 0:15:37.280 --> 0:15:42.280 So consolidation is a concern though in general of AI, it's. 0:15:42.120 --> 0:15:44.640 Certainly something we have to be vigilant about to make 0:15:44.720 --> 0:15:47.920 sure that to the degree we do see you know, 0:15:48.160 --> 0:15:52.520 monopolies or firms with enormous market power, we want to 0:15:52.560 --> 0:15:56.960 make sure a that that's not being achieved through illegal tactics, 0:15:57.800 --> 0:16:00.920 or that it's not being used illegal to snuff out 0:16:00.920 --> 0:16:04.000 competition somewhere else. So that's incredibly important. 0:16:04.840 --> 0:16:07.920 How about the training data situation, because that feels like 0:16:07.960 --> 0:16:10.120 a mixture of different parts of the government be involved, 0:16:10.120 --> 0:16:14.240 but that also feels like a massive anti competitive issue. 0:16:14.320 --> 0:16:16.840 Is this something the FTC is looking into or would 0:16:16.880 --> 0:16:20.720 have to partner with other parts of the government potentially, So. 0:16:20.640 --> 0:16:23.720 We are looking into it. Actually, last year we convened 0:16:23.720 --> 0:16:27.760 a roundtable with creators from all sorts of different sectors. 0:16:27.760 --> 0:16:34.000 So we had you know, artists, authors, graphic designers, even 0:16:34.160 --> 0:16:40.320 people who work in fashion sharing how their core creations 0:16:40.400 --> 0:16:44.200 had actually already been scraped in many instances and suddenly 0:16:44.240 --> 0:16:47.480 they were waking up where their life's creation having been 0:16:47.680 --> 0:16:50.800 ingested by this machine and spitting out outcomes that they 0:16:50.840 --> 0:16:53.840 had never consented to and suddenly felt like they had 0:16:53.840 --> 0:16:56.600 no control over. And so that's something we've heard a 0:16:56.600 --> 0:17:00.680 lot about. Interestingly, these creators, many of them were acknowledging 0:17:00.720 --> 0:17:04.480 that these technologies could be quite fruitful for them as well. 0:17:04.640 --> 0:17:08.720 They're not kind of reflexively and viscerally anti AI. It's 0:17:08.760 --> 0:17:11.919 really about wanting to make sure that the terms on 0:17:12.000 --> 0:17:16.440 which their creations are being used, that those are actually 0:17:16.480 --> 0:17:20.000 fairly negotiated and it's not some type of opt out 0:17:20.080 --> 0:17:23.280 model after the fact that they actually should be able 0:17:23.359 --> 0:17:27.200 to consent, that they should actually be getting compensation that's 0:17:27.320 --> 0:17:29.480 you know, can measure it with the value of what 0:17:29.520 --> 0:17:32.800 they're getting. So that's a lot of concerns we've heard. 0:17:34.040 --> 0:17:37.720 The FTC submitted a comment to the Copyright Office, which 0:17:37.720 --> 0:17:40.639 is considering some of these issues, kind of relaying what 0:17:40.640 --> 0:17:44.240 we had heard, and so it's certainly a live issue, 0:17:44.280 --> 0:17:47.080 and to the extent the FTC can share expertise on it, 0:17:47.119 --> 0:17:48.199 we've been looking to do that. 0:17:49.040 --> 0:17:52.439 So Generati is being crammed into every tech product. Do 0:17:52.480 --> 0:17:55.120 you see any potential for consumer homes from this happening. 0:17:57.680 --> 0:18:01.680 Well, generally, we've brought a set of lawsuits that fall 0:18:01.720 --> 0:18:05.080 into a few categories. One of the big categories is 0:18:05.280 --> 0:18:09.520 around this idea of AI hype. So as more and 0:18:09.560 --> 0:18:14.840 more products are adopting generative AI services, they're making oftentimes 0:18:15.280 --> 0:18:19.840 commitments or promises about what these services can actually deliver, 0:18:20.520 --> 0:18:26.040 and sometimes we see that those claims are inflated. And 0:18:26.119 --> 0:18:29.240 so you know, for example, the FTC a few months 0:18:29.280 --> 0:18:34.240 ago brought a whole sweep of enforcement actions against for example, 0:18:34.280 --> 0:18:37.359 a company that was claiming to be the world's first 0:18:37.440 --> 0:18:41.240 robot lawyer. Do not pay, but the product fail to 0:18:41.280 --> 0:18:44.200 live up to its claims that the service could substitute 0:18:44.240 --> 0:18:47.200 for the expertise of a human lawyer. Now, don't get 0:18:47.200 --> 0:18:49.919 me wrong. To the extent that these types of services 0:18:49.960 --> 0:18:54.600 and advances can bring things like way low cost legal 0:18:54.640 --> 0:18:57.679 services to people, that's a great thing. But just we 0:18:57.800 --> 0:19:01.680 want to make sure companies are not inflating or hyping 0:19:02.000 --> 0:19:06.840 or falsely overstating what those services can actually do. So 0:19:06.920 --> 0:19:09.480 that's been a big strain of the cases that we've 0:19:09.520 --> 0:19:10.760 brought so far in this area. 0:19:12.119 --> 0:19:15.280 So do you see any anti competitive issues in the 0:19:15.280 --> 0:19:18.480 cloud in the compute industry in general, even outside of AI, 0:19:18.920 --> 0:19:21.120 because it really is just a few big players. 0:19:23.480 --> 0:19:26.960 It is. Yeah, it is a relatively small number of firms. 0:19:27.720 --> 0:19:31.359 The FTC did do a inquiry here. We got a 0:19:31.400 --> 0:19:35.000 lot of comments and some of the complaints that we 0:19:35.080 --> 0:19:39.120 heard about focused on things like how easy is it 0:19:39.240 --> 0:19:45.960 to exit the cloud service and transfer your data elsewhere? 0:19:46.440 --> 0:19:50.359 And are you having to play pay these egress fees 0:19:51.520 --> 0:19:55.960 to the cloud providers in ways that inhibits, you know, switching. 0:19:57.280 --> 0:20:00.679 We also heard just more generally about some of the 0:20:00.760 --> 0:20:05.520 software licensing practices, where some people suggested that some of 0:20:05.520 --> 0:20:10.200 the cloud providers limit their ability to use certain software 0:20:10.240 --> 0:20:14.080 in other cloud infrastructure provider environments, so that you can 0:20:14.119 --> 0:20:16.800 imagine could lead to lock in, or if one cloud 0:20:16.800 --> 0:20:21.720 provider is effectively degrading the ability to use other third 0:20:21.760 --> 0:20:25.320 party softwares on that cloud, you could see some issues there. 0:20:26.359 --> 0:20:29.399 And then we also heard some concerns that some provisions 0:20:29.400 --> 0:20:34.040 in cloud computing contracts could be incentivizing customers to consolidate 0:20:34.119 --> 0:20:37.479 their use of cloud services to just one provider that 0:20:37.520 --> 0:20:42.399 could further hasten consolidation. More generally, we looked at this 0:20:42.640 --> 0:20:45.119 not just from the competition lens, but also from a 0:20:45.200 --> 0:20:48.919 resiliency lens. I think we've all seen how when you 0:20:49.040 --> 0:20:53.119 have extreme centralization that can also create a lot of 0:20:53.119 --> 0:20:56.480 fragility where you have a single point of failure, right, 0:20:56.520 --> 0:21:00.240 And we've seen how if a single data of a 0:21:00.640 --> 0:21:04.000 center somewhere goes out, sometimes a corner of the Internet 0:21:04.080 --> 0:21:06.480 also goes out. And so how do we make sure 0:21:06.480 --> 0:21:10.600 we're also thinking about the resiliency element here. And then 0:21:10.600 --> 0:21:13.639 they're kind of similar data security concerns as well that 0:21:13.680 --> 0:21:14.320 we heard about. 0:21:29.000 --> 0:21:33.120 So you've been a nonmeditorializing here attacked I would describe 0:21:33.119 --> 0:21:34.960 some of the clammiest men alive. A lot of people 0:21:35.000 --> 0:21:37.919 in bench capital and within tech have come for you. 0:21:38.000 --> 0:21:40.720 How do you deal with that? How's that affected you? 0:21:44.240 --> 0:21:47.320 So as a general matter, you know, I focus on 0:21:48.280 --> 0:21:52.160 making sure we're hearing from Americans across the country. And 0:21:52.359 --> 0:21:55.760 sometimes it's very easy in these jobs to actually become 0:21:55.880 --> 0:21:59.520 very insulated and only hear from the people who have 0:21:59.640 --> 0:22:02.360 access us to power or kind of know you know, 0:22:03.200 --> 0:22:05.399 where to be ranting in a way that it gets 0:22:05.480 --> 0:22:09.679 to public enforcers very easily. So it's really about just 0:22:09.720 --> 0:22:14.679 making sure that you're keeping good perspective, keeping a broad 0:22:14.760 --> 0:22:18.160 purview of who it is that we serve. So we've 0:22:18.160 --> 0:22:21.080 done listening sessions across the country, for example, going to 0:22:21.160 --> 0:22:26.879 places like Baraboo, Wisconsin or Aimes, Iowa, hearing from you know, 0:22:27.080 --> 0:22:29.679 the gig drivers that use some of these ride sharing 0:22:29.720 --> 0:22:33.879 platforms to understand what are they seeing, what is their experience. 0:22:34.760 --> 0:22:37.000 We talk to a lot of farmers who are concerned 0:22:37.000 --> 0:22:40.560 about the ability to repair their own tractors and why 0:22:40.600 --> 0:22:44.680 there's been you know, increasing friction there and so keeping 0:22:44.840 --> 0:22:48.480 just a broader set of voices in mind as we're 0:22:48.520 --> 0:22:52.600 making our decisions, I think has been incredibly important to 0:22:52.800 --> 0:22:56.399 just keep a north star that's properly focused on, you know, 0:22:56.480 --> 0:23:00.040 the public as a whole, rather than a handful of 0:23:00.040 --> 0:23:04.640 of well resourced, well connected individuals that can sometimes have 0:23:04.680 --> 0:23:07.680 an outsized role in policy decisions. 0:23:09.080 --> 0:23:11.399 So what do you think, what are some of the 0:23:11.440 --> 0:23:14.000 big wins? What the ones you're most proud of in 0:23:14.040 --> 0:23:14.440 your time? 0:23:15.480 --> 0:23:20.280 Well, because of the FTC's work, Americans saw the cost 0:23:20.320 --> 0:23:23.320 of inhalers go down from hundreds of dollars just just 0:23:23.440 --> 0:23:27.639 thirty five dollars. In the digital space, you know, we've 0:23:27.840 --> 0:23:30.760 really helped set new rules of the road. I mean, 0:23:30.800 --> 0:23:33.840 one trend that we've seen is as more companies have 0:23:34.960 --> 0:23:40.439 become reliant on service based revenues, they've introduced subscriptions and 0:23:40.480 --> 0:23:43.679 become much more reliant on subscription revenue, and so we 0:23:44.160 --> 0:23:46.639 see firms make it very easy to sign up for 0:23:46.680 --> 0:23:51.440 a subscription, but then extraordinarily difficult to cancel. We brought 0:23:51.440 --> 0:23:54.840 a whole bunch of lawsuits, including against Amazon, including against 0:23:54.920 --> 0:24:01.000 Adobe for their illegal subscription practices that were awarding Americans 0:24:01.000 --> 0:24:04.560 from being able to cancel in ways that resulted in people, 0:24:04.760 --> 0:24:08.320 you know, overpaying for months and months. We had people 0:24:08.359 --> 0:24:11.000 tell us the only way I could cancel my subscription 0:24:11.160 --> 0:24:13.639 was to cancel my credit card, right, I mean, just 0:24:13.760 --> 0:24:18.320 deeply anti anti consumer practices. We also finalized a rule 0:24:18.560 --> 0:24:21.840 that will go into effect next year such that all 0:24:21.840 --> 0:24:24.199 companies would have to make it as easy to cancel 0:24:24.240 --> 0:24:26.520 a subscription as it is to sign up for one. 0:24:26.920 --> 0:24:29.760 We've also been very active in taking on data brokers. 0:24:30.480 --> 0:24:32.400 You know, there's been a wild wild West out there 0:24:32.440 --> 0:24:36.880 and how people's personal information is bought and sold, and 0:24:37.000 --> 0:24:39.480 we've brought a whole set of law enforcement actions when 0:24:39.640 --> 0:24:44.200 data brokers are illegally using or collecting people's health data. 0:24:45.119 --> 0:24:51.280 People's precise geolocation data, sensitive data like browsing data. Just 0:24:51.320 --> 0:24:54.520 to make clear that the default assumption is not that 0:24:55.000 --> 0:24:57.720 all of this sensitive data can be bought and sold 0:24:58.200 --> 0:25:02.480 even if people have not affirmative given permission. We just 0:25:02.560 --> 0:25:06.320