Meta, YouTube, and a damning addiction verdict

0:00This BBC podcast is supported by ads outside the UK.

0:30to bbc.com and the BBC app. Find out more at bbc.com slash unlimited. Matt, you've had a little bit of drama this week. Are we calling it drama? Well, yes. So, I mean, the first thing is we're both very excited that we're doing Cast Fest, which is a podcast festival run by the BBC. All the best podcasts, like AmeriCast, NewsCast, you name it, and obviously top comment. Actual journalists like Laura Koonsberg are going to be there as well. And us. And us. you can come along to it's in May Deveille it's in London and if you go to the link in the description you can sign up for tickets to come along yes but the drama you're referring to is that my name was at the bottom of the line up and very small on the poster it was like Matt was like the supporting supporting act yeah I still I feel like maybe people at the BBC don't understand how big of a deal I am yet because I have the tiny chair as well tiny name tiny chair like I'm getting but I understand they've now changed the lineup and my name is now quite near the top and much bigger.

1:36So I don't know if my agents had a word with them or what's going on there. Yeah, or we've done a bit of whispering behind the scenes. You probably assumed that it bruised my ego and preempted that. Okay, other than CastFest, we are going to be talking about loads of stuff on the episode today. We're starting with this verdict that's come in, this landmark case that was happening in California. I know landmark is kind of overused, but it actually kind of has been, all about social media addiction and the companies and what they're accountable for and not. This is probably the biggest story to happen so far since we've been doing this podcast in social media.

2:07We're also going to talk about the rise and rise of AI influencers. As in ones that are actually AI generated. Yes, influencers that don't exist.

2:24okay so this verdict that's coming in california it's a really big deal because meta which owns instagram and facebook and then uh youtube which is owned by google have been found liable in this case which is essentially about social media addiction so it's about a 20 year old who goes by the name of kaylee who sued these companies at several of the companies actually some settled some ended up in court saying that the impact of being addicted to social media as a kid massively has impacted her and her mental health and tiktok and snap actually settled before the case began which is why we're focusing on on meta and youtube they argued there's no evidence that the issues that this young woman experienced were because of our platforms um obviously she argued the opposite her lawyers argued the opposite and the jury found them liable which means that she's getting awarded something like$6 million.

3:11So$3 million in compensation and$3 million that's like punitive damages. And the reason this is such a big deal is because this is not just one person. This is what's called a bellwether case. There are over 10 ,000 minors or young adults who allege harm or addiction caused by social media companies. There are over 800 school districts who allege this. And this case was picked because it's representative of a large number of these cases of children and young people being allegedly harmed by social media. She's won. This now changes everything. But what is so interesting about this case and this litigation is the clever way that the prosecution have made this about how the platforms are designed, not the content on the platform.

3:57So until now, loads of people have tried to sue social media companies, but they've always been protected by a law called Section 230, which has also been called the 26 words that created the Internet. It's basically a U.S. law that means that social media companies and Internet-based platforms cannot be held liable for the content people post on those platforms. They're not publishers, that kind of argument. Exactly. This one law has arguably allowed all these social media companies to become the force that shapes our world. But this litigation is so clever because it's not saying the specific content being published is the issue.

4:35Therefore, free speech is not the question here. It's alleging that these platforms are intentionally designed in a way that makes young people addicted. Not only that, but these companies knew that and did it anyways. And that's what's so striking about this. What this jury has come back and said is this is deliberate. They make these design choices and these design choices have an impact on people. It's not happening sort of by accident or as a byproduct. It is a kind of conscious, intentional decision. And that feels really different from anything that's come before. Absolutely. And it's because one pro-anorexia video, for example, in itself is arguably defensible by free speech and isn't that harmful.

5:16But if a platform is designed in theory to send 10 ,000 pro-anorexia videos to one person, the way that platform designed, they allege, is intentionally harmful. And the companies obviously disagree with this verdict. They've made that very clear. Meta said, quote, teen mental health is profoundly complex and cannot be linked to a single app. We will continue to defend ourselves rigorously as every case is different. And we remain confident in our record of protecting teens online. And then the spokesperson for Google said, this case misunderstands YouTube, which is a responsibly built streaming platform, not a social media site.

5:52And both of them are going to appeal this case. And obviously, you imagine that they've got one eye on the other cases that might come. The other 9 ,999 press cases. That could end up the same way or not. I mean, it really reminds me of that thing we've spoken about before, which is how the companies have often argued, we are a mirror. Like, we're holding up a mirror to society. We don't actually affect people. It's just we're reflecting it back to you. And essentially what this case says is that is not true. These are proactive systems. They are features that you've designed and implemented and they're having an impact.

6:23And this comes at around the same time that Meta was ordered to pay$375 million in damages in a New Mexico case for allegedly misleading users about the safety of its platform. This is possibly the worst week for social media companies ever. Zuck is not happy this week. I actually did an exclusive interview with Anna Lemke, author of Dopamine Nation, who was an expert witness in the California trial. And she's also a world leading addiction psychiatrist. And this is what she said about the verdict. I didn't know that the public was ready to identify just how addictive social media can be. And my worry was that so many people themselves use social media that they would want to normalize it and not look at the problematic piece of it.

7:16Adults are not obviously immune to this problem. And I think that probably did impact the jury verdict. I suspect that in describing these phenomenon, the jury saw themselves in this behavior, you know, and it resonated. They said, yeah, this is true. I've had this experience on these platforms. You know, early on, people were likening it to big tobacco litigation. I think that's a fair analogy. But I had the sense that maybe we would have to lose a bunch of cases before we would win them. But it's clear to me that we have reached a tipping point, that even with these very beginning trials, we're getting guilty verdicts, we're getting, you know, large punitive damages.

7:57This is going to change things for these corporations. To me, it's really interesting thinking about how a jury would respond in this case, because I guess the thing about social media is that it is just so universal. And yet it increasingly feels like, and maybe this is just the state of my DMs and why I feel like this, that people are much more aware, much more vigilant as to what they're being pushed, why they're being pushed, how the platforms actually work. And so it strikes me that someone like Anna, who knows loads about this, was kind of like, I'm not sure whether people are going to realise that this is the impact that the sites are having.

8:29And what seems to have happened is actually they've heard the evidence and thought, oh, I kind of recognize this. Yeah, because although these cases are about young people and children, this applies to adults, too. This applies to all of us. And when Anna Lemke was explaining how these platforms are designed with something called persuasive design to keep us on them as long as possible, it resonated so much with me. I began to realize that I was addicted, too. So they have neuroscientists, the top neuroscientists in the world, working on this. This stuff comes from the world of gambling, of slot machines.

9:02These are features like infinite scroll, autoplay. There's also recommendation algorithms, follower counts, which give you dopamine because the perception that you are being liked by a social group becomes addictive. The sense of FOMO, push notifications. These are all intentionally built in because they know exactly how to keep you hooked and how to control your dopamine reward cycle in your brain. But where it gets quite neuroscience-y, and I think quite sinister, is in this concept of novelty, uncertainty, and variable rewards. So the way our brain works, the way dopamine in our brain works, is a constant dopamine stream, a constant reward, actually keeps us less hooked than variable.

9:50Here's a bit, now we take it away. here's a bit now we take it away so they use this novelty uncertainty principle where they intentionally do not let the algorithm give you exactly what you want all the time because that actually keeps you going back for more rather than getting what you want all the time then you kind of feel sated and you're tolerant to the dopamine so they they give you a bit of dopamine then they stop bit of dopamine stop and you don't want to become almost like desensitized to that level of They don't want you to. They don't want you to. You want to. The idea that you are always about to get the thing that you want is apparently one of the most potent ways to keep us hooked.

10:26So when you think of a slot machine, if you need three lemons and it'll do one lemon, then it'll give you another lemon. You nearly get it. Then a cherry. Yeah, yeah, yeah. That is how these algorithms are designed. We know that intermittent rewards are actually more addictive than reliable rewards. So when those rewards come at an unpredictable interval, or we're not quite sure when we're going to get the reward, we're more likely to stay engaged in anticipation of that unknown reward. You know, these companies have incorporated that kind of slot machine design knowingly into their product such that you have this pull to refresh phenomenon or the quote unquote mystery of the algorithm, their language, such that part of what keeps us engaged is this kind of treasure seeking function, not knowing, you know, what's around the next algorithmic bend.

11:16If you become tolerant to the same kind of reward, they then switch it up and give you a different version of that. She compared this to if you're addicted to cookies and someone's like shoving a chocolate chip cookie in your face constantly. And then you're like, OK, I don't want any more chocolate chip cookies. That's when they would switch it. And now it's a chocolate chip cookie with almonds. Yeah. So it's different. You're mixing it up a bit. You're like, oh, OK, I kind of this is a this is a variation on a theme. Exactly. I quite like. And the novelty uncertainty thing would be like, actually, I'm not going to give you any more chocolate chip cookies.

11:45I'm just going to give you oranges and apples. And you're like, oh. And then you get a cookie again and you're like, yes. Exactly. And when I talk to people who work within the social media companies, especially engineers, I'd say, so those are the people basically in charge of designing and controlling these algorithmic systems. I mean, they talk about that, that ultimately they are optimizing for engagement, whatever the cost. And I guess the best way of optimizing for engagement is to operate this kind of reward-withdraw type system. You brought up earlier this argument that they've often made that we're just holding a mirror up to you you are the ones making the content this is what you are creating and what you want to see we're just you know we're just posting it but this is so different from what we've heard in this trial not only not holding up a mirror to us but i mean things like fomo we've all felt fomo when looking at social media but to know that they arguably are aware of fomo and are intentionally putting it in the product to give that to you.

12:44There's like a neuroscientist sat there being like, more FOMO. I mean, it's terrifying. I imagine that's live scenes, live scenes from any social media headquarters you can think of. But this is the point, and actually I think this is where, when we think about the consequences of this case and the other cases, what they could be, people within the companies, they tell me, the biggest worry is that they are asked, not necessarily in the United States because of the current sort of political environment and the approach to the companies, but certainly in other places, they're basically asked to overhaul their models.

13:16They're asked to overhaul their entire algorithmic structures and how they operate, which is what makes them all the money. That's the thing. If they've then got to have more proactive moderation, slower algorithms, they've got to separate out the platforms that kids are on and teenagers are on to the main ones, all of that is incredibly expensive to execute. And actually, it's the FOMO and the keeping you scrolling that is hugely valuable to them, which is why they've continued to build on that rather than the opposite. If you're working inside a social media company this week and you support what they're doing, you'd be thinking, well, do we risk being regulated in a much harsher way that's much more difficult for us to deal with in some places?

13:55Do we risk getting banned in much more places? And also the public perception point, which might explain why some of the companies settled and others didn't. Because now it's kind of like Meta and in this case, Meta and YouTube and big sort of bold lights are the ones who, if you use social media, you might be looking and thinking, oh, maybe I don't want to use those platforms then if that's the case. I think there is a tipping point. I think people are waking up to this and the perception of social media as a bad thing is growing really, really fast. I think people are comparing this to big tobacco now.

14:26Another aspect of big tobacco is they can't advertise to adults either. So, you know, we're focusing on children now for good reason, but there are huge questions to be raised about, should they be allowed to do this to us? Digital media is a legal drug. I mean, it's the Wild West, but it's also a legal drug. And we as a society have determined that we can't market addictive drugs to kids. So if we extend the analogy, then we have a perfect argument for saying, gee whiz, maybe kids shouldn't be allowed to go on these addictive platforms. One of the things that really came out in this litigation is that there are lots of tweens and even younger children on defendants' platforms, and the platforms know that they're there and are mostly doing nothing about it.

15:10I don't think there's ever going to be a day when we make digital drugs illegal. But I do think that there needs to be much more transparency, more warning. People need to really appreciate that the digital media is a powerful tool, but also potentially a potent drug. It is worth us saying, obviously, that there's a chance that that the other cases don't go in the same direction as this one, that it is very much case by case. It's a test case, not class action here. And so the companies, and you'd really notice this in Meta's response to this case, they are like, it's case by case. We've got to see what happens.

15:49And, you know, as someone who investigates social media, I find, I don't know if you find this, Matt, there are real examples where people properly evidence the impact that social media has had on them. There are also examples of where people point to harm and it's not quite what it seems. Obviously, you'd imagine that, you know, when a case gets as far as being in court like this, that there's pretty significant evidence to back it up. But that's not to say that it's all going to completely go in this direction. It might end up being not quite so straightforward as that. Yeah, and I think one can imagine that Zuck and the others are horrified by the prospect that it does, because imagine$6 million in damages, you know, times 10 ,000.

16:26And it's also interesting to note that Meta, in this case, owe 70 % of the damages. YouTube. More than YouTube, yeah. The other 30%. So YouTube perhaps was considered less addictive and problematic in this case. Yeah, and it's kind of interesting, the idea that the social media platforms vary in terms of the possible harm that they might be able to cause to you. I mean, YouTube in particular, when I was doing the investigation about algorithms, I looked into YouTube as I did all the big companies, but in terms of the evidence to back up like what's happening right now and making decisions that exacerbate harm or exacerbate the problem, There was lots of evidence when it came to several of the other companies, but actually YouTube in some ways, and they highlight this in their response, have tried to move from being kind of old rabbit hole YouTube, you know, where people would end up like goodness knows where, into we're like a streamer.

17:14We're like, you know, one of the we're like a video platform. That's not to say there's not problems. Of course there are. But you also have to remember that a lot of these cases are about the experiences that teenagers had growing up. Yeah, I mean, I think this is clearly going to have huge ramifications for how social media companies who are mainly based in the US where this litigation is happening deal with children. Age verification will potentially have to become more stringent. Algorithms will potentially have to be rewritten. Or you just operate completely separate platforms altogether.

17:42Like you literally have to have platforms potentially different spaces. I mean, it could mean social media in the way that we have known it for kind of our whole lives is a really different thing, certainly for people in like five, 10 years time. That does raise the question of like, do they have to entirely rebuild their platforms? Like, do they have to start from scratch? I mean, lots of people talk about safety by design. The problem is, is that the designs don't feel inherently safe. And a lot of the engineers I speak to say it's kind of a black box. It doesn't feel like we can fix the algorithms very easily.

18:14And that's sort of why we're here. That excuse may no longer hold water now that this litigation has happened. This also has ramifications for AI companies and whether their content is harmful for children. And ultimately, it sets a precedent in which tech companies in general can be held liable for the way their products are designed and whether those designs are harmful and addictive, not just for young people, for all of us. So this could change everything.

18:51So AI influencers have very much now become a thing, like capital A, capital T. You might have seen them popping up on your feed. They've just become kind of much more widespread. And perhaps that's because of the advancements in terms of the AI tech and how people can make them. And they're increasingly convincing. They look like real people, real influencers. And it's not just that they look like influencers, They're also kind of treated like real people, like influencers. There's literally been this AI Influencers Awards that have been launched this week, possibly as part of a bid to kind of incentivise people to create these kinds of characters and this kind of content.

19:29But the main reason that we're talking about this isn't the awards. It's actually because there is one particular AI influencer who has been going super, super viral. She is the so-called MAGA Dream Girl. So who is she? So she is called... Or it? It makes me think of like the AI agents rebelling when we call them it. The clankers. Yeah, it freaks me out. Do you know, by the way, just a quick side note on clankers, clankers is a slur for robots and AI. I bet that's never been said, by the way, in any BBC podcast studio ever. Is that because of... A quick aside on clankers. Well, like on Blue Sky, which is like the super, super woke version of Twitter, to the point where it's almost comical, there's like a debate over whether using the term clankers to refer to AI and robots is actually offensive.

20:11A slur. AI-ism. Yeah, exactly. So I just think let's be mindful of calling them clankers. Yeah, let's be mindful. We don't want to get like... Okay, so she's a clanker. So yeah, she's called... Maga Dream Girl is called Jessica Foster. You might have seen her popping up. I mean, the reason that she's been popping up recently is obviously because there is this war going on in the Middle East and lots of the images that Jessica Foster has popped up in are ones where she's wearing like an army uniform. She's blonde. I mean, she's absolutely beautiful as are all these AI influencers, which is kind of the point.

20:42And she's really promoting a kind of certain conservative image or ideology. But she's more of an OnlyFans model. But she's more of an OnlyFans model. Like her whole thing is basically being gorgeous. And like posing next to Donald Trump and F-22 Raptor jets. Yeah, yeah, yeah. And you'll notice that her feet prominently feature in a lot of her posts because for reasons unknown to mankind, one of the largest corners of the internet is the foot fetishist corner. and that's like a guaranteed way to grow an audience and make money. Yeah, and now Jessica Foster's account has actually been taken down from Instagram, also from OnlyFans for violating their policies, presumably on synthetic material, because it's not always clear that she's not real.

21:25And I think the motive here, Matt, is so interesting because it's like, it reinforces certain political ideas. But also, she had over a million followers in four months on Instagram. I mean, that is a huge tool that if you want to like make money, you want to monetize your content, you want to, I mean, particularly direct people to your OnlyFans where you can make lots of money, you can see why someone would do it. It's both a business and then I guess it's more of a business because it's political. But it's got to be not just a hot girl, but she has to fit a certain archetype. So this is Maga Dream Girl.

21:56There's also a trucker version. There's a trucker girl. There's hot cop girls as well. But not all of these AI influencers are just for money. So I've noticed, I don't know if you've seen this, loads of AI-generated, attractive Iranian female soldiers. Oh, I had seen some of this popping up. Now, do you know what the biggest giveaway that these aren't real Iranian female soldiers is? Tell me. Iran doesn't allow women to serve in combat rules. What's interesting about these Iranian AI-influencer female soldiers is they are speaking in English. They're not wearing the hijab. So that kind of indicates that they are aimed at English speaking audiences, which I think is interesting because it's a lot of a lot of Iran's war slop that it's coming out with right now.

22:41And it is winning the war as far as war slop is concerned, is aimed towards English people. So that is an example of how these influencers aren't just for money. They're also for political reasons, too. Danny Bones. Oh, yeah. So Danny Bones is a kind of working class British rapper with a fast growing online following. He has millions of views. He wraps up immigration. In one of his rap videos, he kind of is leading like men carrying St. George's cross. He's also wearing a jersey that says mass deportation unit. He's not real. He's made he's an AI musician. and the Bureau of Investigative Journalism discovered that he is a front for an anonymous influencer collective called the Node Project.

23:29The Node Project was paid by the Advance UK party. Which is like a really quite fringe far-right party. Yeah, Advance UK paid the Node Project to make its main campaign video and repurpose some of its videos for its Gorton and Denton by-election. I think the really crucial thing here is, do people care that these people aren't real? And in fact, most of the time, they probably know that they're not real. But isn't this the point that when it comes to influencers in general, I mean, the word influencer is almost synonymous to some extent with the word fake. And that's not to be derogatory by influencers, but it's just this perception that you are trying to influence people.

24:04Therefore, you've come in with an agenda. You're often paid to promote something. You're an ad, whatever. And whether you're a real person or whether you're an AI generated person, I mean, the effect is kind of the same. It's sort of like fake on steroids. so you can see why people are actually just like well i don't really mind if this person isn't real because they fulfill this image in my mind or something that i agree with politically or that i like and actually that's what's so powerful about ai is that it can kind of leave an impression quite a strong impression of something and it doesn't really matter whether it's real or not it starts to kind of shape how you view reality i do i do think that most people would care that it isn't a real person for whatever reason i feel like we need to feel like it's a real human being that we are quite connecting connecting with yeah yeah maybe not so much when it's like an only fans creator and it's just people like who are horny and don't care but when it's someone like danny bones or like a musician um you know we we want to feel like there's a real human being that we're connecting with and when it comes to politics i mean we've spoken quite a lot on various episodes about authenticity and politicians and actually how the content that performs best when it comes to politicians and leaders is the stuff that seems genuine, seems authentic.

25:20It's kind of funny in some ways that this political content and these AI influences are very much the opposite of authentic. And yet they kind of succeed in pushing this hyper-stylized message. It's the too good to be true thing, right? You see this stuff and you think, oh, I really want this to be true. I'd like the world to be like this. I like this person. I find this person attractive. whatever it is and it's like well it is too good to be true but i just feel like it's becoming it's being really normalized and so actually everyone's kind of getting to the point where it becomes a sort of fair game thing to do whether it's to make money through it or to actually use it to further an agenda i mean you can get an ai girlfriend or boyfriend now i love how you said that like i was You could.

26:03Are you okay? Now, okay, what are the rules on this? So TikTok, yes, AI influencers are allowed on TikTok, but they must adhere to strict transparency guidelines, including clearly labeling their content as AI-generated. Yeah, so mandatory disclosure. They cannot impersonate other people with AI. Meta not only allows, but actively supports AI influencers and creators on Facebook through its AI studio program. So this stuff gets so much engagement, right? Like it's good for the creators. It's good for the platforms in that sense. But there is a darker side to this as well. I mean, aside from the whole like tampering with reality point.