WEBVTT - FTC Chair Lina Khan Talks Future of AI Regulation

0:00:02.440 --> 0:00:06.800
<v Speaker 1>Bloomberg Audio Studios, podcasts, radio news.

0:00:07.160 --> 0:00:08.960
<v Speaker 2>You just got out of speaking with the folks at

0:00:09.080 --> 0:00:11.719
<v Speaker 2>Y Combinator. Talk to us a little bit about who

0:00:11.800 --> 0:00:13.560
<v Speaker 2>you're speaking to and why you're here.

0:00:13.600 --> 0:00:14.440
<v Speaker 1>What's the message.

0:00:15.280 --> 0:00:17.720
<v Speaker 3>Well, thanks so much for having me, Emily, And an

0:00:17.720 --> 0:00:21.000
<v Speaker 3>important part of my job is going around the country

0:00:21.160 --> 0:00:25.880
<v Speaker 3>and hearing from Americans across markets, across sectors to understand

0:00:25.920 --> 0:00:28.400
<v Speaker 3>what are the pain points that they're seeing in the

0:00:28.440 --> 0:00:32.400
<v Speaker 3>markets that they're trying to work in, compete in, buy in.

0:00:32.920 --> 0:00:34.680
<v Speaker 3>And so I'm so glad to be here in San

0:00:34.680 --> 0:00:39.480
<v Speaker 3>Francisco meeting with folks across the tech community, including founders,

0:00:39.560 --> 0:00:43.839
<v Speaker 3>including vcs, including startups, and trying to understand, you know,

0:00:43.920 --> 0:00:46.680
<v Speaker 3>how do we make sure that our free and fair

0:00:47.040 --> 0:00:52.040
<v Speaker 3>enterprise system can remain robust and that innovation and competition

0:00:52.120 --> 0:00:56.080
<v Speaker 3>can thrive, especially at this moment of technological change and

0:00:56.200 --> 0:00:57.600
<v Speaker 3>a potential inflection point.

0:00:58.600 --> 0:01:01.120
<v Speaker 2>Obviously, how AI will play out is very much top

0:01:01.160 --> 0:01:04.160
<v Speaker 2>of mind here. I just spoke with Meta CEO Mark Zuckerberg,

0:01:04.160 --> 0:01:08.400
<v Speaker 2>who unveiled the largest open source AI model ever this week,

0:01:08.440 --> 0:01:10.920
<v Speaker 2>and really double down on what he thinks are the

0:01:10.959 --> 0:01:13.800
<v Speaker 2>benefits of the open source approach. Take a listen to

0:01:13.840 --> 0:01:14.600
<v Speaker 2>what he had to say.

0:01:15.440 --> 0:01:17.600
<v Speaker 1>The way that we can control our own destiny on

0:01:17.640 --> 0:01:19.440
<v Speaker 1>this and make sure that we have access to leading

0:01:19.480 --> 0:01:23.119
<v Speaker 1>AI is by building it and having it become an

0:01:23.120 --> 0:01:26.720
<v Speaker 1>industry's standard. But it actually gets stronger by being able

0:01:26.720 --> 0:01:29.080
<v Speaker 1>to share it and have the ecosystem around it.

0:01:31.000 --> 0:01:34.560
<v Speaker 2>Where do you stand cher con on Meta's open approach

0:01:34.720 --> 0:01:37.680
<v Speaker 2>versus closed source competitors like open AI and Google and

0:01:38.040 --> 0:01:42.240
<v Speaker 2>Zuckerberg's remark specifically, well, look, without.

0:01:42.400 --> 0:01:45.559
<v Speaker 3>Weighing in on any specific company. As a general matter,

0:01:45.800 --> 0:01:49.640
<v Speaker 3>historically we've seen that openness and open source can be

0:01:49.720 --> 0:01:52.920
<v Speaker 3>a critical vector of innovation, right. I mean, what happens

0:01:52.920 --> 0:01:56.559
<v Speaker 3>traditionally when you render some of these ecosystems more open

0:01:56.960 --> 0:01:59.720
<v Speaker 3>is that it can lower barriers to entry, It can

0:01:59.760 --> 0:02:04.560
<v Speaker 3>lower costs for entrepreneurs and startups and really enable them

0:02:04.600 --> 0:02:07.600
<v Speaker 3>to innovate much more freely. And so we want to

0:02:07.600 --> 0:02:12.440
<v Speaker 3>make sure that that tradition continues. We are watching very closely.

0:02:12.480 --> 0:02:15.679
<v Speaker 3>I think there's still an open question around what openness

0:02:15.720 --> 0:02:19.200
<v Speaker 3>really means in the AI context. Our team has been

0:02:19.200 --> 0:02:23.160
<v Speaker 3>focusing on this idea of open weight models, which we

0:02:23.280 --> 0:02:27.160
<v Speaker 3>think could similarly engender a lot of competition and innovation

0:02:27.360 --> 0:02:30.840
<v Speaker 3>and make it easier for startups and entrepreneurs to compete

0:02:30.919 --> 0:02:33.520
<v Speaker 3>on a level playing field. Of course, we need to

0:02:33.800 --> 0:02:38.800
<v Speaker 3>really closely understand what are any restrictions licensing restrictions that

0:02:38.840 --> 0:02:42.120
<v Speaker 3>are being included here and could those be fore closing competition.

0:02:42.560 --> 0:02:46.239
<v Speaker 3>More generally, we also need to be aware of how

0:02:46.280 --> 0:02:50.679
<v Speaker 3>there are significant incumbents that may have an outsized role

0:02:50.720 --> 0:02:54.160
<v Speaker 3>to be playing because they have key control over the

0:02:54.240 --> 0:02:57.280
<v Speaker 3>raw material and key inputs, and so we need to

0:02:57.280 --> 0:03:00.440
<v Speaker 3>be aware of the broader potential competition issue us or

0:03:00.440 --> 0:03:03.760
<v Speaker 3>bottlenecks or choke points that could emerge in ways that

0:03:03.800 --> 0:03:07.400
<v Speaker 3>could inhibit innovation and inhibit competition. And so that's what

0:03:07.440 --> 0:03:09.800
<v Speaker 3>we're going to be talking with founders and vcs and

0:03:09.840 --> 0:03:10.760
<v Speaker 3>others about today.

0:03:12.040 --> 0:03:14.799
<v Speaker 2>Something else you've been looking into more recently is surveillance pricing,

0:03:14.800 --> 0:03:18.520
<v Speaker 2>which is how AI is being used to rapidly change

0:03:18.520 --> 0:03:21.000
<v Speaker 2>pricing based on consumer behavior. And there are some really

0:03:21.040 --> 0:03:24.520
<v Speaker 2>dystopian concerns here about how this could lead to dangerous

0:03:24.840 --> 0:03:28.280
<v Speaker 2>financial targeting. What are some examples you've seen of this

0:03:28.400 --> 0:03:29.959
<v Speaker 2>and what are you hoping to find.

0:03:31.520 --> 0:03:34.880
<v Speaker 3>So as you know, for years now, Americans have had

0:03:34.960 --> 0:03:39.240
<v Speaker 3>their personal data and information closely tracked and surveilled by

0:03:39.280 --> 0:03:42.040
<v Speaker 3>a whole set of companies in ways that can really

0:03:42.080 --> 0:03:45.760
<v Speaker 3>threaten people's privacy. What we see now is that firms

0:03:45.800 --> 0:03:50.200
<v Speaker 3>could also be using this enormously personal data about people

0:03:50.360 --> 0:03:54.480
<v Speaker 3>to also set person specific prices based on what they

0:03:54.520 --> 0:03:57.560
<v Speaker 3>know about you, based on what you're feeling, who you're seeing,

0:03:57.760 --> 0:04:00.640
<v Speaker 3>what you're browsing, and so we want to make sure

0:04:00.760 --> 0:04:03.480
<v Speaker 3>that we are looking under the hood and understand what's

0:04:03.520 --> 0:04:06.760
<v Speaker 3>going on here. So, for example, you can imagine that

0:04:07.040 --> 0:04:11.040
<v Speaker 3>a hotel that knows you've already bought plane tickets may

0:04:11.120 --> 0:04:14.040
<v Speaker 3>show you a higher price than they show somebody who

0:04:14.080 --> 0:04:17.240
<v Speaker 3>hasn't really decided on where they're going to visit. You

0:04:17.279 --> 0:04:21.120
<v Speaker 3>can imagine a restaurant that knows you have a one

0:04:21.240 --> 0:04:26.359
<v Speaker 3>hour lunch break could hike prices for you during that hour.

0:04:26.440 --> 0:04:28.080
<v Speaker 3>And so these are just some of the types of

0:04:28.120 --> 0:04:31.400
<v Speaker 3>practices that we could see take hold if this type

0:04:31.440 --> 0:04:34.039
<v Speaker 3>of surveillance pricing is allowed to develop.

0:04:34.920 --> 0:04:38.359
<v Speaker 2>Interesting we're all still feeling the effects of this crowd

0:04:38.400 --> 0:04:41.159
<v Speaker 2>strike outage, and so many people are shocked about how

0:04:41.200 --> 0:04:43.760
<v Speaker 2>a company that is so small could just shake up

0:04:44.120 --> 0:04:47.080
<v Speaker 2>the world like this. What's your take on this and

0:04:47.160 --> 0:04:48.720
<v Speaker 2>is there something to investigate here.

0:04:50.920 --> 0:04:53.800
<v Speaker 3>So as a general matter, I think this really underscores

0:04:53.920 --> 0:04:58.120
<v Speaker 3>the trade offs between kind of fragility and resiliency, and

0:04:58.160 --> 0:05:01.000
<v Speaker 3>how when you have more centralization, or when you have

0:05:01.320 --> 0:05:04.640
<v Speaker 3>enormous dependence on just one firm or a very small

0:05:04.839 --> 0:05:07.599
<v Speaker 3>number of companies, that can make it so that a

0:05:07.720 --> 0:05:11.919
<v Speaker 3>single disruption can have cascading effects and really lead to

0:05:11.960 --> 0:05:14.760
<v Speaker 3>all sorts of breakdowns. And so as a result of

0:05:14.760 --> 0:05:18.279
<v Speaker 3>this one incident, for example, we saw over three thousand

0:05:18.320 --> 0:05:22.080
<v Speaker 3>flights grounded. We saw doctors that were no longer able

0:05:22.160 --> 0:05:27.200
<v Speaker 3>to prescribe medicines for their patients, We saw courthouses that

0:05:27.279 --> 0:05:30.680
<v Speaker 3>had to close off proceedings, and so the impact here

0:05:30.880 --> 0:05:34.400
<v Speaker 3>was significant, and it's really underscoring the importance of making

0:05:34.480 --> 0:05:38.200
<v Speaker 3>sure we have resilient markets and resilient systems so that

0:05:38.279 --> 0:05:41.440
<v Speaker 3>a single crash like this doesn't lead to so much

0:05:41.480 --> 0:05:42.560
<v Speaker 3>outsized disruption.

0:05:43.880 --> 0:05:45.400
<v Speaker 2>I do have to take a moment to talk about

0:05:45.400 --> 0:05:48.800
<v Speaker 2>the presidential election. A number of prominent tech folks have

0:05:48.920 --> 0:05:52.080
<v Speaker 2>endorsed President Trump for a variety of reasons, but I

0:05:52.120 --> 0:05:56.640
<v Speaker 2>have heard specifically multiple tech executives and investors complain about

0:05:56.640 --> 0:06:00.240
<v Speaker 2>how they can't do deals on your watch, you have

0:06:00.279 --> 0:06:04.599
<v Speaker 2>any concerns that your agenda has that all alienated the

0:06:04.680 --> 0:06:10.120
<v Speaker 2>tech community or could impact the impact of your agenda

0:06:10.240 --> 0:06:10.919
<v Speaker 2>going forward.

0:06:12.560 --> 0:06:14.359
<v Speaker 3>You know, it's been such an honor to serve and

0:06:14.400 --> 0:06:17.520
<v Speaker 3>then Biden Harris administration, and you know, we're just focused

0:06:17.520 --> 0:06:21.359
<v Speaker 3>on doing our work. What I oftentimes hear from the

0:06:21.360 --> 0:06:26.520
<v Speaker 3>business community, including small businesses, including entrepreneurs, is that they

0:06:26.600 --> 0:06:29.839
<v Speaker 3>want markets to be more open and more fair and

0:06:29.960 --> 0:06:34.240
<v Speaker 3>more competitive, rather than incumbents being able to squash out

0:06:34.400 --> 0:06:37.880
<v Speaker 3>maason competitive threats. I mean, our free enterprise system is

0:06:37.920 --> 0:06:40.840
<v Speaker 3>one where the best ideas are supposed to win, and

0:06:40.880 --> 0:06:44.560
<v Speaker 3>we've historically seen that it's the disruptors and entrepreneurs that

0:06:44.640 --> 0:06:47.320
<v Speaker 3>have been a key vector of innovation. And so our

0:06:47.440 --> 0:06:50.320
<v Speaker 3>job at the FTC when we enforce the antitrust laws

0:06:50.360 --> 0:06:53.000
<v Speaker 3>is to make sure that our markets stay open and

0:06:53.040 --> 0:06:56.719
<v Speaker 3>fair and competitive. And that's something that you know, most

0:06:56.760 --> 0:07:00.440
<v Speaker 3>businesses and most entrepreneurs should really be able to get behind.

0:07:02.000 --> 0:07:05.760
<v Speaker 2>Former President Trump's running mate jd Vance has actually praised

0:07:05.880 --> 0:07:08.920
<v Speaker 2>your work. In fact, I believe he said that you're

0:07:08.960 --> 0:07:11.120
<v Speaker 2>one of the few people in the Biden administration that's

0:07:11.160 --> 0:07:14.640
<v Speaker 2>doing a pretty good job. In his opinion, what do

0:07:14.680 --> 0:07:18.120
<v Speaker 2>you make of the irony there? And you know, how

0:07:18.200 --> 0:07:21.520
<v Speaker 2>do you see the FTC moving forward in the next administration.

0:07:23.160 --> 0:07:25.920
<v Speaker 3>So I will speculate about what's going to happen in November.

0:07:26.560 --> 0:07:30.520
<v Speaker 3>I do think we see enormous bipartisan agreement that we

0:07:30.600 --> 0:07:34.040
<v Speaker 3>need to reinvigorate and double down on anti trust enforcement.

0:07:34.560 --> 0:07:37.760
<v Speaker 3>We need to make sure that monopolies can't illegally use

0:07:37.800 --> 0:07:42.000
<v Speaker 3>their power to jack up prices for consumers, to really

0:07:42.120 --> 0:07:46.880
<v Speaker 3>undermine workers, or really stamp out entrepreneurs in small businesses.

0:07:46.920 --> 0:07:50.800
<v Speaker 3>And so that bipart is an agreement around the importance

0:07:50.880 --> 0:07:54.600
<v Speaker 3>of open, fair competition to protect our free and fair

0:07:54.760 --> 0:07:58.160
<v Speaker 3>enterprise system is absolutely critical. I mean, you know, just

0:07:58.240 --> 0:08:01.760
<v Speaker 3>the other week, the FTC release an interim staff report

0:08:02.080 --> 0:08:05.520
<v Speaker 3>looking at the practices of the pharmacy benefit managers and

0:08:05.560 --> 0:08:09.120
<v Speaker 3>some of our initial findings. And that's an area, for example,

0:08:09.120 --> 0:08:12.760
<v Speaker 3>where we see enormous bipartisan concern on both sides.