Why is everything a conspiracy?

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

0:29CFO. LinkedIn has a word for that. Bull spend. Now you can invest in what looks good to your CFO. LinkedIn ads generates the highest ROAS of all major ad networks. You'll reach the right buyers because you can target by company, industry, job title, and more. So cut the bull spend. Advertise on LinkedIn. The network that works for you. Spend$250 on your first campaign on LinkedIn ads and get a 250 credit for the next one. Just go to linkedin.com slash broadcast. That's linkedin.com slash broadcast. Terms and conditions apply. Your social media feed delivers plenty of advice, but it doesn't know you.

1:14It doesn't ask questions. It doesn't give physical exams or order tests. Doctors do. At the American Medical Association, we believe the best care starts with a real conversation with someone who understands the science and your unique health. So stay curious, ask questions, but when it's time to make decisions, make them with a doctor. Learn more at amahealthvshype.org. That's amahealthvshype.org.

1:44If they could make us all get fired and hire metabots instead of people, that would be a huge victory and they are trying to bring that world into fruition. Yes, they're coming for you. Hello and welcome back to The Interface, the show that decodes the tech that's rewiring your week and your world. I'm Thomas Germain. I'm Karen Howe. I'm Nikki Wolf. Today on The Interface, we'll be looking at how the Trump assassination attempt reveals the conspiracy theory playbook. Meta is spying on its employees to replace them with AI, and you might be next. And AI is making our world a little too perfect.

2:24So there's a new app you can use to add typos to your emails.

2:33So I was in Washington, D.C. this weekend, and it was a bit of a strange weekend. I don't know if you guys saw the news. Yes. It was the White House Correspondents' Dinner this weekend, which is this big annual gala for the White House Press Corps. There's the main event, which I wasn't at. There was a load of satellite parties, there was a party nearby, with actually a load of people who used to work in the national security industry. And everyone's phone went off all at the same time. And it was because, and I'm sure you've seen this, there was an attempted shooting at the White House Correspondence Dinner.

3:09Now, we're still finding out about what exactly happened kind of as we go. This was only a couple of days ago as we record this. But that has not stopped the internet from taking what happened and running with it. It's formed almost a perfect storm of conspiracy theory happenings in a way that I think is very, very illustrative of how this kind of stuff often happens. So the second I heard that this had happened, you know, you can feel the gears start turning. There's going to be some people at least who are out there telling some crazy story. but what do we know about what actually happened this weekend?

3:47So this guy called Cole Allen, who was still learning about exactly who he is, but he ran through the security at the correspondence dinner. Shots were fired. Some shots were fired by him, it seems. Some were fired by the Secret Service. He was taken into custody on the scene. once the shooter was named, people quickly started digging online and found an anonymous Twitter account that in 2023 posted one tweet. This is the only tweet that this account has posted. And it just says, Cole Allen, the name of the shooter. Unarguably, that's weird. Yeah, it's pretty weird. So the conspiracy theory came about that a time traveler was warning us about this attack.

4:47That makes sense. What? Sorry, a time traveler posted the tweet. What part of this is confusing? The time traveler went back to 2023 to warn people about this shooting via the medium of the... if I was a time traveler going back to 2023, there would be things I would do first, I think. Yeah, well, I would just have more of a plan. You know, if I got my hands on time travel technology, I wouldn't get back. I was like, wow, I didn't think about this. I guess I'll tweet this guy's name and hope people catch on to it. Like, they could have given us some more clues. The thing is with this stuff, and like for anyone who's a conspiracy theorist, like this is probably enraging and like we're part of it obviously because we're picking this apart.

5:38But like, could you walk us through step-by-step the conspiracy playbook? Because you can see it the same way every time, right? This isn't some kind of special situation. It's just another conspiracy theory. So first, while it's still happening, you get the misinformation about what's actually going on. You get theories that there'll be a second gunman. You get theories that there's unexploded bombs. people are sharing stuff who aren't on the scene, people are seeing stuff that's being shared from the scene that isn't necessarily from a trustworthy source. Then after the thing is over, immediately you'll start to get misidentification.

6:15So this didn't actually happen in the White House correspondence sense, but you will often get like a suspect who's named who everyone will find on the kind of old Facebook account or something and everyone will file. Like if we don't know who did it, Like Sluice will go try and identify someone off and they'll pick the wrong guy. Yeah. And then the authorities will name a suspect and then people will take that name and go hog wild with it. So in this case, this is how you got people finding that old time traveler-y tweet. They dig through the social media. Then you get people in finer and finer detail going through all of the videos and going through the information.

6:54That's the playbook, basically. That's the kind of thing to expect. And it almost always brews on the same subreddits. Yeah, it'll be happening on the same subreddits. You will also see, so there's kind of below Reddit level, there's a site called 4Champs, I think we've talked about before. These are kind of fully anonymous forums. And in some instances on these forums, you will get people actively trying to seed conspiracies. Not saying that's what happened with this time traveler tweet, but a lot of the time you will have people making a Photoshop kind of fake picture of something like that, like something from the past, and then spreading it.

7:35And because lots of people are on these sites, they can seed it in a way that will then make its way into the short-form video algorithmic feed ecosystem. So I'm still wondering, like, why does this always happen? The thing with this stuff, and this is a term that came out of QAnon, which was this massive kind of conglomeration of conspiracy theories that I did a lot of reporting on. There's this idea of baking, which people who go into QAnon were called bakers because they were following the breadcrumbs. And people love following little clues. It's kind of empowering, you know, unraveling conspiracy theories, right?

8:12There's all these like massive global forces that are acting upon us and the public all the time. Especially in a world where you feel so powerless. Right. If you can figure stuff out, it's like you're on the inside. you know what's really happening. You're catching the bad guys. And the other is that, especially on the internet, it's a way to generate attention and make money, right? There's this whole ecosystem that profits off this kind of stuff, right? Yeah. What would happen if Reddit just shut down the main subreddits that usually brew these conspiracy theories? Do you think that that would significantly ameliorate the problem or would it actually just push this activity to other parts of the internet?

8:56This is not a factor of any particular part of the internet. It is a fundamental part of human nature. And the reason to not kind of shut this down too much is that actually, ultimately, it is not necessarily unhealthy to not trust everything that you are told by people in authority, right? So that is not in and of itself a bad instinct. AI is going to make this kind of thing much harder. That's going to throw a spanner in all these words. It's easier to produce this information quicker. Yeah. Well, I think the thing to keep in mind here is way more nefarious and convoluted conspiracy theories than this one have happened and turn out to be real.

9:48Like, it's not like this sort of thing is unimaginable. But every single time, every time, without exception, that something bad and shocking happens, there will be all these conspiracy theories picking it apart and telling you about how it's some kind of master plan and nothing is real. So I think it's worth understanding what the tropes are because if you start to notice, well, this is just the same thing every time, that's a useful tool to help sort out whether things are actually truly bizarre or whether just like every real world event, there's something odd if you look too closely at it.

10:28Yeah.

10:33Go on, get a little out there into the big heart of Nevada where you can go off-road and off the map on two lakes or on horseback. Dip into hot springs and dive into deserts. Climb a mountain or make your best effort. See thousands of stars in some of the darkest skies. Stake out haunted hotels. Can you make it to sunrise? There's always something new to see because we've got plenty of space to just be. Plan your trip at travelnevada.com. This is Jana Kramer from Windown with Jana Kramer. Instead of giving your mom something that fades, give her something that becomes part of her home this Mother's Day.

11:12The Lennox Spice Village is a set of 24 hand-painted little houses that are actually spice jars, perfect for anyone who loves to cook, entertain, or enjoy the little details that make everyday life special. As a mom, I love gifts that help turn ordinary moments into memories. Charming, timeless, and meant to be used, this is one of those pieces she'll treasure. And once you see it, you'll want it for your own home too. Find the full collection at lennox.com slash spice village. Owning a home is full of surprises. Some wonderful, some not so much. And when something breaks, it can feel like the whole day unravels.

11:50That's why HomeServe exists. For as little as$4.99 a month, you'll always have someone to call. A trusted professional ready to help. Bringing peace of mind to 4.5 million homeowners nationwide. For plans starting at just$4.99 a month, go to homeserve.com. That's homeserve.com. Not available everywhere. Most plans range between$4.99 to$11.99 a month your first year. Terms apply on covered repairs.

12:14So I have to say, I am not going to bring a palate cleanser to that story. I'm going to talk about more scary things, but not as scary as an attempted assassination. Yes. So there was this scoop from Reuters that broke that Meta is installing new tracking software onto all of its employees' devices for tracking their clicking, their mouse movements, occasionally taking screenshots and other things related to their work-related activity. And this is all ultimately for the purpose of trying to improve their AI agents to perform work-related tasks. This is a story that intersects so many different trends that are happening in the world simultaneously.

13:02The first one being that, of course, everyone has heard, AI might take your job. And in this particular instance, Meta is not only really orienting, like they're really explicitly coming out here and being like, we are on the path to trying to develop labor automating software. where the first people that we are going to try to automate is our employees. And the second people might be you because they are training these agents to do the work in-house and then to potentially sell it as a service or just release it open source into the world for other companies to do the exact same thing. But the other thing that's kind of an interesting trend that this hits upon is this general feeling that surveillance and AI is also just taking more and more agency and power away from workers in general and giving it back to the capital owners, the business owners.

14:03And obviously, this has been a controversial announcement to employees for various different reasons. But one of them is simply that they cannot opt out. Workplace surveillance is already pretty common. Workplace tracking is pretty common. And especially in the tech industry, it's pretty common. But this is just a degree of invasiveness that I think really surprised meta employees, especially because the executives made clear during an employee all hands that if they don't like it, that means they just have to leave the company or apparently move to Europe because Europe actually has data privacy protection.

14:48They have laws. They have laws over there. Must be nice. We're a little bit screwed, but... Just so I'm clear. So every single thing that a Meta employee does in their computer now, like every mouse movement, everything they type is like being logged and recorded. Not everything. So what the internal memos have said and what the executives have said during these employee all hands is that it's specifically work-related apps, but Gmail is considered a work-related app. So one of the concerns that employees brought up is that What about when we're just checking our Gmail for, you know, like, to talk to our doctor, you know, like something that's like actually not work related, but happens to be on the work related device and in a work related application.

15:35And the executives basically were like, well, then just don't do it on your work laptop. Like you should do that elsewhere. So yeah, so it is supposed to be targeting very specific types of data. But obviously, it's going to be hard to really only get work-related information. Yeah. So Reuters reached out to Meta and a Meta spokesperson did mention that they have safeguards to protect employee-sensitive content. They didn't actually elaborate on what those safeguards are. They actually also didn't elaborate that to employees either during the employee all-hands. It's just like, don't worry about it.

16:17Yeah, but what's kind of interesting is the spokesperson then just reiterated essentially what we've been talking about here, which is if we're building agents to help people complete everyday tasks using computers, our models need real examples of how people actually use them. So they're just kind of reinforcing, yes, they're coming for you. Why would an AI agent being trained need to learn on stuff like mouse clicks, right? AI is not interacting with a computer using a mouse. How does that help? This is a really interesting question because there's an answer for why this is being attempted, but there's also caveats to it.

16:58So the answer is like AI agents are ultimately navigating in a world, a digital world that's designed for humans. And so some of the things that Meta wants to capture is like clicking buttons in software or navigating a drop-down venue, things that humans would typically have to do because that is how software has been designed. And AI agents aren't actually just interacting at a behind-the-scenes level. They have to interact at this user interface level. But some people have brought up this question of, is this actually the right way to train AI agents in the first place? Like, they could actually just communicate more fundamentally with software like ingesting HTML code, which is the code that tells the computer how to visualize a website.

17:48Like, instead of using the pixels to capture the information on the website, like, why not just ingest the code directly? So this is both a valid question and a kind of strange approach. But potentially, it's because these agents aren't going to have access to the underlying code of a piece of software all the time. You know, they can't just, let's say you have Microsoft Word downloaded on your laptop. Like, the likelihood that Microsoft as a company would be willing to open up their code directly to a meta AI agent. is like low probability. And so the AI agent will probably need to navigate different pieces of software that are owned by different companies just through that visual interface.

18:35So actually, the app could get around things like signal encryption. Say someone's got the Signal app on a work phone, it could open it in the way that the user would on the screen rather than having to get into the encrypted actual app. That's terrifying. Yes, yes, exactly. The thing that strikes me so much about this particular scoop is that the AI industry for a while was really trying to make themselves appear good. You know, like they're this moral force. They're trying to ultimately work on things like scientific discovery and improving health care and education and whatever. And no, of course not.

19:17We're not trying to automate jobs. And now it's just all of that is melted away. And it's like, yeah, we're coming after our own employees jobs first to use their data to then come after your jobs. Like it's it's we've gotten to the, you know, the evolution or the maturation of the industry to where they they simply don't have to hide anymore what the actual main agenda is. And they're just going after it full steam ahead. Yeah, like if you're a meta employee, like what is happening here is like they are spying on you to try and create a little robot guy who ideally for the company does what you do, essentially.

19:58And then there's this, you know, fear, right? We've heard often from the CEOs of AI companies that there will be like mass unemployment because these tools will get so good that they'll do all of our jobs and we'll all be on the street begging robots for change or something. I guess part of the question here is like, I think people hearing that obviously will be alarmed, right? They're like specifically to create a white collar office worker. Like, how successful do you think these efforts are going to be? Will they, like, are jobs going to be replaced because of this thing? Or is it not as a, do you expect it won't be as effective as, you know, the most dire predictions suggest?

20:43The first thing to say is projects like this have been tried before. The monitoring of mouse movements, keystrokes, screenshots to train AI agents, and they have not actually worked. One of the hypotheses about why it might be different this time is because of the sheer scale. Like, Meta has a lot of employees and no one's allowed to opt out. So are they going to be able to capture, you know, usually doesn't work because the data is really noisy and there's a small sample size. But is somehow the scale going to help overcome some of those challenges? But this is a really legitimate caveat to this whole enterprise is that is Meta actually going to successfully create the white collar worker?

21:31Probably not, because it turns out that a lot of work is really complicated. You know, Darren Asimoglu, this institute professor at MIT who won the Nobel Prize for Economics, he always points out that in Silicon Valley, there's this ideology that all aspects of work are kind of computable. And in the end, robots would actually do every aspect of work better than a human as long as we just train it on the right data. But when you look at the reality of production, there's a lot of things that are not computable. Like there's a lot of social emotional relationship, change management, like leadership skills involved in things and like dealing with surprising occurrences, dealing with, you know, just like the messiness of the world.

22:25And it turns out that actually AI automated systems are really not that good at that. Humans are particularly good at that. And so there is this big question of even with a sheer amount of scaled data for doing mouse movements. Yeah, sure. Maybe the AI agent will finally learn how to click a bunch of buttons and navigate dropdown menus. But is that sufficient enough to truly replace all forms of work? And Darren Osamu's answer is no, it's not. But what's interesting, that does not mean that there won't be potentially big layoffs in the near future due to the orientation of the AI industry in this direction.

23:10I mean, this scoop came out simultaneously with Meta also publicly announcing that they are laying off 8 ,000 employees and are going to leave 6 ,000 job openings vacant. They are not going to actually hire for them. And I think most people would assume, oh my God, Meta is successfully automating all this work away. and that is why they're not hiring these employees. But it's actually a little bit more nuanced than that. They are making a future bet that based on their current investments, they will be able to eventually automate that work away. And they kind of need, they're under a lot of pressure to cut costs now.

23:52So they're just gonna lay off those workers and not hire those workers immediately. And the reason they're under that pressure is because they're spending$135 billion this year on building out AI infrastructure. This is kind of a little case study for how other companies are going to operate. A lot of other companies are going to see the AI industry orienting towards trying to automate more and more work. And they, as consumers of those AI technologies, are then going to think, huh, well, maybe I can also make a future bet that eventually these AI tools are going to help me automate away most of the work that my employees are doing right now.

24:32But I kind of need to cut costs now to be more competitive. So let me just do the layoffs and slow down hiring in the present. Well, the Wall Street Journal just reported that in March alone, the tech industry laid off 45 ,800 workers, right? Like layoffs, layoffs, layoffs, we've been hearing every month, another company, and not just in the tech industry, like, you know, across other parts of the economy too. I think it is fair to say in the tech industry that layoffs are happening because of AI. But it isn't necessarily because jobs are being replaced by AI. It's because they are spending so much money invested, like an amount of money that like it's so large that the human mind cannot wrap itself around numbers this big.

25:21That's how many dollars are being spent on this stuff. And on top of that, we have just exited this period where, especially for the tech industry, money was free, right? Interest rates were really low. Advertising on the internet was just like, it's like they had their own mint and they were just making cash. Everything has tightened up. And now they have to live in the real world. And they were probably really overstaffed. and people are, you know, being let go because the companies now suddenly have to act like real businesses and not like infinite money machines. Yeah, obviously this is pretty confusing.

25:57There's some things, should we be worried about losing our jobs because the AI is too good? Should we be worried about losing our jobs because the companies think it's too good, but it's not? Should we be worried about privacy because they're trying to create, you know, should we be worried? Karen, help us sort through this. We need a guide here. Yes, we should be worried. Okay. Oh, good. Okay. Okay, great. But we're not sure exactly what we should be worried about. We don't know how worried you should be. Have a baseline level of worry. What you should be worried about. You know, I mean, they're trying to create a tool that will do people's jobs ostensibly because that will be cheaper than a real person and you'll pay Meta instead of paying an employee if you're some random company.

26:42It is not at all a given that this will succeed, that it will be able to do your job and that this is going to be a huge problem. But it is certainly something to pay attention to. And it's worth understanding that that is the goal of this industry. Like if they could make us all get fired and hire metabots instead of people, that would be a huge victory. And they are trying to bring that world into fruition. And of course, this topic is hugely important and very complex. So we will definitely return to the topic of how AI is going to impact work and labor in the future. All right, let's shift away from the world of work here.