Lina Khan is a different kind of Federal Trade Commission chair. Khan, who is only 34, made her name with a much-discussed paper titled “Amazon’s Antitrust Paradox” while she was still in Yale Law School. The paper, which bucked longstanding consensus about how to evaluate monopolies, marked her as a leading voice against corporate consolidation — a stance that has largely been out of fashion in Washington for decades. During her two years leading the FTC, Khan has pursued an aggressive regulatory strategy in sectors from big tech to book publishing, with tactics that have inspired many admirers and many critics. In the latest episode of On With Kara Swisher, Kara speaks with Khan about how to regulate new tools like ChatGPT, her focus on Meta in particular, and, of course, Elon Musk.
On With Kara Swisher
Kara Swisher: You just wrote a terrific op-ed in the New York Times, calling for more regulation on AI. Explain your argument and why it’s not too late.
Lina Khan: The argument was really underscoring that we have all of these new AI tools entering the market that are becoming more widely adopted, but there’s no AI exemption from the laws already on the books. Sometimes there can be a perception that these new tools are coming amid this regulatory vacuum, and my argument was that actually, we already have laws and tools, and we don’t even necessarily need new ones. Maybe we do, and Congress can decide that. But let’s not forget that the existing laws, the existing rules prohibiting deception, prohibiting unfairness, prohibiting discrimination, prohibiting collusion — all of those laws still apply, and where the FTC has authority, we’re going to be vigorously enforcing those laws, even if it involves AI tools.
Swisher: People seem to love playing with these chatbots, kind of like they did with social media. It’s fun, it’s fascinating. The dangers can seem abstract and sci-fi-ish. A lot of polls show people aren’t that worried, even though the people who make this stuff are more worried. Walk us through two of the practical downsides that you wrote about: scam and fraud.
Khan: Because of our jurisdiction, we are entirely focused on the here and now. And we’ve already seen how these AI tools, while they might provide a lot of opportunity and benefits, can also be misused in ways that are turbocharging fraud. They allow scammers and fraudsters to disseminate fake content much more cheaply, much more quickly, and on a much wider scale. And so we’ve already seen, for example, how people have used voice cloning to make phone calls, to pretend to be people’s family members, to do very, very convincing impersonations, pretend to be a family member in distress and end up scamming family members out of thousands of dollars. We’ve seen how some of these tools can be used to generate very, very convincing phishing attempts, where you’re using somebody’s social media profile to basically generate a message that is entirely targeted to them and what their vulnerabilities might be. And again, we’re seeing people being scammed out of thousands of dollars.
Swisher: And in this case, it’s not linked to Section 230, which gave the early internet companies broad immunity. These are actually the same companies, but they don’t get that exemption here, correct?
Khan: That’s right. I mean, I’m sure there’ll be various issues tested in the courts, but our sense is that Section 230 is unlikely to apply here. The companies that created these tools could be on the hook.
Swisher: And they can’t say they’re just a benign platform, from your perspective.
Khan: That’s right.
Swisher: When I interviewed Sam Altman, who’s the CEO of Open AI, he essentially said the safest way to deal with AI is to release it to the public and fix issues as they come up. Yet the FTC recently warned AI firms not to over rely on “post-release detection.” You’d like them to test and catch in advance. Is it possible to anticipate all the issues?
Khan: I think we need to be very wary of systems where it’s the public that is carrying the burden of doing cleanup. I mean, these companies have the information, they have the resources. These are multibillion-dollar firms, and the responsibility should really be on them to make sure that before they put these tools out into the wild, that they’re safe for people.I think a system where you’re putting the burden on, under-resourced civil society groups, under-resourced academics, under-resourced public servants, is just going to create a system where you’re endlessly doing cleanup rather than addressing things on the front end by the actors that are best positioned to do that.
Swisher: So was this a warning to them?
Khan: Absolutely. We’re looking to make sure that businesses are on notice, that the FTC is watching, that our existing laws and authorities apply in this space, and that they need to be complying with the law as they introduce these tools.
Swisher: Another thing the FTC is worried about is AI bias. You put out a joint statement with the DOJ’s Civil Rights Division and the Consumer Financial Protection Board. I love a statement just as much as anyone else, but what’s the point?
Khan: The point is to make sure these firms know that the existing laws prohibiting discrimination — each of our agencies has laws, be it the Civil Rights laws, the Equal Credit Opportunity Act — that these laws are going to apply. And so if their tools are being used to automate discrimination, they need to be on guard to make sure they know that these laws prohibiting that discrimination are going to be applied. And they could be on the hook for that.
Swisher: Do you think it scares them in any way?
Khan: I think it’s really important for us as enforcers to, again, make sure that the market knows that these existing tools apply. I think we’ve seen all too often instances in which companies play fast and loose with the rules, especially right at the beginning when there can be some confusion about, “Hey, are these new tools? Did the existing laws apply? What does that look like?”
Swisher: “What, me? Huh?”
Khan: Exactly So we as enforcers thought it was incredibly important for there to be absolutely no doubt, for there to be clarity and transparency for businesses, that they know all of these existing laws apply and we as law enforcers are going to be applying them.
Swisher: Do you think that makes them nervous or wary or not? Some people suggest the power to regulate AI should lie with the FTC, but you are understaffed as it is.
Khan: I think the big issue here is a lot of these tools are just going to be used in ways that are naturally part of our work. We go after fraud and scams. When these tools are being used to, to propagate fraud and scams on a massive scale. That’s clearly within our jurisdiction. We enforce the nation’s antitrust laws, so when we see mergers or acquisitions or potentially collusive behavior that is being facilitated through these tools, that’s also going to be part of our work.
So I think this is a broader point where increasingly the distinction between digital and non-digital or tech and non-tech is being blurred as these tools and as these technologies get implemented across business. So if we get a merger of grocery stores, there can be a really significant digital aspect of that merger. Relating to how these companies are using the data. So I would just say the digital aspect of our work is increasing and expanding across the board, just given the realities of how businesses are evolving.
Swisher: Okay, let’s move on to Facebook. What’s the philosophical underpinning for the monopoly case against them?
Khan: Facebook is a repeat violator at the FTC. There was a consent decree that goes back close to a decade, which the FTC in 2019 found that they violated. The recent news suggests that they may have also been in violation of this latest consent order. And that is really prompting a step back and a close look at: What does it take to make sure that firms across the board are actually complying with the law? Are fines, even if they’re big, enough? Or do we as an agency need to be looking at the underlying business models and the incentives that those business models are creating to make sure that firms are not able just to treat these consent decrees or even fines as a cost of doing business, and that they’re actually fully complying with the law.
Swisher: So in this case, where does it stand right now?
Khan: So we voted out an order that would change a consent decree. Now Facebook has an opportunity to make a filing, and then we’ll take it from there. We also have a separate lawsuit that was filed before I arrived at the FTC, looking to address the fact that Facebook’s acquisitions of Instagram and WhatsApp were illegal — that it allowed them, basically, to maintain its monopoly as the market was transitioning from desktop to mobile. So that lawsuit is underway. When I came in, the judge had tossed the case out. We filed an amended complaint. We survived a motion to dismiss, and so now the trial is proceeding.
Swisher: Does Khan have a Meta fixation? That was in the Wall Street Journal. It seems like the Wall Street Journal has a Lina Khan fixation, but …
Khan: I think when you have companies that are repeatedly before a law-enforcement agency, you need to ask serious questions about whether these companies are recidivist and whether they have a challenge in abiding by existing laws.
Swisher: Recidivist — that’s a good word. That’s a loaded word. Meaning that they will continue to do this until you stop them.
Khan: I think we have a serious problem where companies — oftentimes big companies — can treat law-enforcement actions or even fines as a cost to doing business. And we need to make sure that they’re abiding by the law and that honest businesses that are abiding by the law are not at a competitive disadvantage because they’re actually following the rules, and losing out to companies that aren’t.
Swisher: Do you think Meta is particularly a particularly bad recidivist?
Khan: The FTC’s own history shows that time and time again, the agency — again, under my predecessors — has had to take action against this company. And so that’s just part of the public record.
Swisher: Twitter is under two consent decrees, speaking of which. In an active investigation, I know you can’t comment on it, but let me pose a hypothetical. Is it okay for a company to misrepresent who is and isn’t paying for the service they offer? For example, I did not buy a blue check, and it looks like I did.
Khan: On its face, that type of deception could be quite concerning and clearly fall within the FTCs jurisdiction.
Swisher: And what are you going to do about it for me particularly?
Khan: [Laughs.] Look, we’ve been following a lot of the developments at Twitter quite closely. As you noted, this is a company that’s been under a consent decree at the FTC for a decade, right? This far preceded any new ownership that took place last year. Even a decade ago, the FTC found that the company was misrepresenting how it was using people’s data. There’s a lot of personal data that’s tied up in Twitter. So anytime we see reporting that that data might be susceptible to misuse, we take that seriously and look at that very closely.
Swisher: Do you think the blue-check thing is misrepresentation?
Khan: There’s a legal test that you go through to show deception, but on its face, I think if there’s a marker that somebody has paid for something, but in fact they haven’t, that’s a concern. I think we also can see glitches and changes where it might not be intentional. So we need to be aware of that distinction as well.
Swisher: According to the GOP’s subcommittee on the weaponization of federal government, you’ve been harassing Twitter. How do you respond to that?
Khan: Again, Twitter has been under FTC order for over a decade. This is a company that was found early on to be playing fast and loose with people’s data. Last year, again, it was found that Twitter had actually violated that original consent decree. We take law enforcement very seriously. We think it’s incredibly important that companies are abiding by the laws on the books.
The consent order is public. It lays out what Twitter was required to do. One of the things it was required to do is monitor and police very closely what types of third parties had access to user data. So if we at the FTC are suddenly reading that all sorts of people have been granted access to Twitter systems, that legal teams that were responsible for complying with the FTC consent order have been fired, it would be malpractice for us to turn a blind eye to that. And so I think it’s absolutely our obligation, especially when you have these repeat offenders, to look very closely to make sure we’re protecting people.
Swisher: So you’re not harassing Twitter?
Khan: Absolutely not.
Swisher: Now, Elon Musk reportedly tried to meet with you. Why did you say no?
Khan: We oftentimes hear from executives about wanting to do meetings, and we consult with our staff who are doing the investigation on the ground. One thing we want to avoid is companies that are obstructing FTC investigation, trying to short-circuit around that by just trying to ask for a meeting at the top. And so we engaged with our staff and learned from them that unfortunately there were still some deficiencies in Twitter’s compliance, and let them know that they should focus their efforts on complying with the FTC’s requests.
Swisher: So did you want to meet with him?
Khan: Look, there’s always a time to meet top-level executives as part of our process. And so if that comes to bear, we’ll be happy to do so.
Swisher: But for right now, no.
Khan: For right now, we recommended that Twitter focused on complying with existing information and requests that FTC staff had put out.
Swisher: So Elon: If you comply, she’ll meet with you, I guess.
This interview has been edited for length and clarity.
On With Kara Swisher is produced by Nayeema Raza, Blakeney Schick, Cristian Castro Rossel, and Megan Burney, with mixing by Fernando Arruda, engineering by Christopher Shurtleff, and theme music by Trackademics. New episodes will drop every Monday and Thursday. Follow the show on Apple Podcasts, Spotify, or wherever you get your podcasts.
More From 'on with kara swisher'
- How Diia Is Becoming the Ukrainian Government’s Real Everything App
- Lilly Singh on Internet Stardom and the Pressures of Representation
- Why Elon Musk’s Twitter Is Like Yahoo Mail