Could AI chatbots counter conspiracy theories?

Could AI chatbots counter conspiracy theories?
If people believe in conspiracy theories, how likely are they to think again when presented with facts and evidence? The good news is that people can shift their view when wacky theories are debunked according to a study just published in the journal Science. The researchers created an AI chatbot that engaged with people and presented them with facts to counter their dodgy beliefs. The study team invited people to state why they believed a conspiracy theory, and fed those responses into GPT4. The chatbot was instructed to be a persuasive sceptic. Thomas Costello, Assistant Professor of Psychology at American University in Washington DC is one of the authors and is on the show.

Potential ban of Chinese EV’s in the US
The US is giving notice on Chinese and Russian manufactured ‘smart cars’ – it’s talking seriously about proposing to ban sales or imports of such connected vehicles. That comes from a White House announcement this week. It’s not a trade war, says the US, but a matter of national security. Gareth and Ania discuss the proposed new rules. 

Original source of migrants eating pets fake news story
“In Springfield, they’re eating the dogs, the people that came in, they’re eating the cats, they’re eating the pets of the people that live there…” said Donald Trump in the televised Presidential debate… Where did such an extreme claim come from? Jack Brewster, Enterprise Editor at NewsGuard found the original comment with his colleague. We hear how he did it. 

The programme is presented by Gareth Mitchell and the studio expert is Ania Lichtarowicz.  

More on this week's stories:
Durably reducing conspiracy beliefs through dialogues with AI
Biden wants to ban Chinese software in 'smart cars' over security concerns
How false claims of pet-eating immigrants caught on 

Transcript

00:00:00 Gareth Mitchell 

Hello everybody, it's Gareth. Welcome along to the Somewhere on Earth podcast from London and it is Tuesday the 1st of October 2024. 

00:00:17 Gareth Mitchell 

And joining me today for some experting, even though Ania doesn't like me calling her an expert, but you’re producer, editing, something like that Ania, people know who you are by now.  

00:00:29 Gareth Mitchell 

But, in what guise are you joining us? Cause I know that you, and quite rightly you like to say that, like Ghislaine and Wairimu and Angelica, you know, they're the experts. I think you're being a bit modest, but what is your function? 

00:00:42 Ania Lichtarowicz 

Very, very kind of you. Well, I think my function today, Gareth actually is a is a bit of a Jack of all trades cuz let's be honest, the interviews have been done by me because you've been rather busy, but that's fine. I forgot how much I enjoyed actually asking the questions. 

00:00:58 Gareth Mitchell 

Yeah, I'm getting worried now because you're exceedingly good at it. So I think my my days here are numbered. So I'll try and be good. So. OK. So Jacqueline of all trades, Ania, as we'll call you for this edition. And when I talk about our stellar lineup of experts, of course, Peter Guest, as if. 

00:01:18 Ania Lichtarowicz 

You forgot him, didn't you? Forgot, Pete. 

00:01:19 Gareth Mitchell 

Forget. I've just saved him till last because well, here we are. 

00:01:24 Gareth Mitchell 

Right. So I think we've done enough damage in this little intro. Let's jump in. Here we go. 

00:01:33 Gareth Mitchell 

And coming up today. 

00:01:38 Gareth Mitchell 

How do you counter conspiracy theories? Well, how about this for an idea? We're going to explore this idea in this edition. How about getting a chat bot to debug these conspiracy theories? We'll get into that. Also, smart vehicles get geopolitical. Yes, not long after having imposed tariffs on electric vehicles from China, the US now says that certain smart cars are potentially a threat to national security. New restrictions could be on their way.  

It's all right here on the Somewhere on Earth podcast. 

00:02:19 Gareth Mitchell 

Right then. Now if you believe in conspiracy theories, as if any of you really do, because you listen to this podcast and you're sensible. But for people who do believe in conspiracy theories, how likely are they to think again 

00:02:33 Gareth Mitchell 

when presented with, you know like facts and evidence. Well, the good news is that people can shift their view when wacky theories are debunked. So says a study just published in the journal Science. The research has created an AI chat bot that engaged with people and presented them with 

00:02:50 Gareth Mitchell 

facts to counter their dodgy beliefs. The study team invited people to state why they might believe a particular conspiracy theory, and then they fed those responses into GPT 4. You know, that very well known chat bot. And it was instructed to be a persuasive skeptic, if you like. Ania has been hearing more from co-author Thomas Costello, 

00:03:11 Gareth Mitchell 

assistant professor of psychology at American University in Washington, DC, and Thomas began by talking about the kinds of conspiracy theories that came up. 

00:03:22 Thomas Costello 

Because people were able to describe their own, that there were a lot, there was, there's a lot of heterogeneity, you know, so some very distinctive conspiracies that that few other people in this sample talked about other than that one person. But there were some notable clusters. So there were things you'd expect like 9/11, 

00:03:41 Thomas Costello 

conspiracy theories. We were doing this on Americans. So JFK assassination was another one. Also things related to COVID-19 vaccinations. 

00:03:52 Thomas Costello 

And more broadly, healthcare companies and other kinds of companies hiding information like cancer cures from the public. There were conspiracies about the 2020 presidential election, so election fraud conspiracies. There were conspiracies about Princess Diana's death, about Martin Luther King Junior's assassination, about the New World  

00:04:12 Thomas Costello 

Order, of course, and related secret societies. Some people talked about celebrities being satanists and things like that. There was one guy who talked about a Loch Ness Monster in a lake in Montana. 

00:04:26 Thomas Costello 

Aliens was another really big cluster. I can't believe I forgot about aliens. So it was really, it was quite a mixed bag, which was one of the cool parts of the study, I think. 

00:04:39 Thomas Costello 

Until AI had been integrated into this kind of experimental paradigm of debunking conspiracies, which was sort of our innovation, you had to plan everything out beforehand. So you had to pick a conspiracy to focus on. And then, of course, not everyone you bring into the lab actually believes that conspiracy. So it's not really a perfect match. 

00:04:59 Thomas Costello 

Whereas by letting people talk about their own conspiracies, we were able to really tailor the evidence and facts that the model provided right to them, which was pretty neat I thought. 

00:05:09 Ania Lichtarowicz 

Wow, I mean that is a really long list you've just given me, including some that I had I had never heard of. And it must be so time consuming to train the AI to be able to argue. Would that be the right word? How did you do that? 

00:05:27 Thomas Costello 

We were able to piggyback a lot off of the work that had already been done by companies like Open AI, whose model we use, but other like kind of AI companies also do this. You know, they fine-tuned and trained the models to follow instructions, but also to adhere to truth and not spread misinformation. 

00:05:49 Thomas Costello 

And so my suspicion, and I've now talked to some of the folks in these AI labs, and I think this suspicion is true, is that they had already kind of like made sure that the models were decent at arguing about conspiracies, and then we were able to just kind of pick up where they left off and finish that job. 

00:06:05 Ania Lichtarowicz 

And then the people came into the lab again and sat down and interacted with the AI. Explain what happened? 

00:06:14 Thomas Costello 

Sure. So the structure of the conversations varied a little bit from person to person, but it did follow a larger overarching pattern. So usually the AI would offer some kind of statement that it understood where they were coming from and that it could see the rationale for why they believed the conspiracy. You know, it wasn't judgmental. But very quickly it transitioned to providing arguments and they were largely factual or rational arguments for why the conspiracy was wrong. 

00:06:46 Thomas Costello 

So in the case of a 9/11 conspiracy theorist who might say that, why did the Twin Towers collapse and burn when jet fuel does not burn hot enough to melt the steel beams of the Twin Towers, which is sort of a famous one. The model would say something like  

00:07:04 Thomas Costello 

yes, that's true, but the steel does have a weakening point that's only about half as hot a temperature required to do that. So what the jet fuel did was it weakened the steel and then everything on top of that steel collapsed and it kind of caused a cascading reaction. And so it's understandable why you thought this, but actually it's wrong. 

00:07:24 Thomas Costello 

And for conspiracies where there was less clear new updated evidence it would make an argument to plausibility. It would say things like this would require tens of thousands of people to be keeping a secret. And when we look at real conspiracies, like again in the case of the US, we had Watergate that was discovered almost immediately, like the real conspiracies get found out pretty quickly, actually. So it's it's not very plausible that all of these people would keep a secret for this many years. 

00:07:54 Ania Lichtarowicz 

And how did the people respond to what sounds to me like quite a rational argument? 

00:08:00 Thomas Costello 

Yeah, they, you know they worked the like wild eyed tinfoil hat wearing conspiracy theorists of a metaphor and kind of cliche. People generally were happy to update their beliefs and in many cases like gratefully. There's one person who comes to mind who  

00:08:20 Thomas Costello 

almost effusively simply thanked the model. There was a minority of folks who became  angry with it and who continued to argue back. So it varied depending on the person, but on average, because the model was so polite, people tended not to get mad. 

00:08:36 Ania Lichtarowicz 

I was going to say, you know, you take out that element of emotion from the argument, don't you with this AI? 

00:08:42 Thomas Costello 

Right. Exactly. You know, people, people argue about controversial things for lots of different reasons and not all of those reasons are about truth or evidence. But when you take the human out of the equation and it's just an AI sharing facts with you, that makes the focal point of the interaction about truth. And perhaps that's why people were open minded in this context. 

00:09:06 Ania Lichtarowicz 

And what about, you know, the impact it had? What was the percentage that people kind of believed the conspiracies less? Was that, that's not a very elegant question, but do you know what I mean? 

00:09:14 Thomas Costello 

How big an effect?  

00:09:18 Ania Lichtarowicz 

How effective was it? Yeah. 

00:09:20 Thomas Costello 

So for the kinds of social science interventions that that this might have a comparison to, you know, psychology studies and political science the effects were really massive, frankly. Like it was almost shocking to see how big a decrease it was relative to other kinds of studies like this. But in an absolute sense. So people went down about 20%. If we use  100 point scale, where 100 is I definitely believe the conspiracy, and 50 is I'm not so sure,  

00:09:48 Thomas Costello 

and zero is I definitely don't believe it. If you started at 80, so you were pretty sure, a 20% decrease is going down to like 60, 66, 67 or so. So it's becoming more sceptical, but it's not totally disavowing the belief, although actually one in four people who started out over 50 ended up 

00:10:10 Thomas Costello 

under 50. And that includes some people who started pretty high. So in some cases people were I guess you could say debunked, but on average, people just became more sceptical. 

00:10:19 Ania Lichtarowicz 

OK. And how long does this last because there's no point doing this if it's going to be a week, but if it's longer? 

00:10:25 Thomas Costello 

Right. So that's why we were so keen to follow up to look at the durability of the effects. We followed up 10 days later and two months later and the effects seemed very long lasting. So people didn't go down at all,  

00:10:40 Thomas Costello 

even two months later relative to where they were right after the conversation. So if you were at that 67 from 80 when we followed up two months later, most likely you'd still be around 67 on that 100 point scale of belief. 

00:10:55 Ania Lichtarowicz 

Would an extra session be beneficial do you think? 

00:10:58 Thomas Costello 

Good question. I mean, that's one of the things that we're looking at now I wish I had an answer, but I mean it seems like probably right. 

00:11:07 Ania Lichtarowicz 

It certainly does, but how you know, how practical is this? Could you adapt it and put it on all the big social media platforms. 

00:11:14 Thomas Costello 

If, if, the social media platforms wanted to do that, it might have an effect immediately, but then there are these, you know, it's entering this complex system, information ecosystem, so people might recoil from it or reduce their trust in the social media companies  

00:11:32 Thomas Costello 

so that the net effect was negative. I imagine we'd want to do a pretty comprehensive analysis of like the ramifications of introducing something like this. And also ethically you don't want to force an argument on people if they don't consent to it outside of a research context where we do get consent. And so I think being thoughtful about like, how to brand this and how to prompt the models to behave is necessary before we like open it up to millions of people. 

00:12:03 Gareth Mitchell 

That was Thomas Costello talking to the excellent Ania. So quite a lot to get into here, really. I mean, for one thing, it was interesting Ania getting his perspective on the kinds of conspiracy theories. So a few that you might have expected the 9/11 theories, Loch Ness Monster and what have you. But were you taken by, either way, surprised or not surprised by the kinds of conspiracy theories that came into this study. 

00:12:29 Ania Lichtarowicz 

Well, I was most surprised by the Loch Ness Monster, to be honest with you. Yeah. I mean, yeah, a lot of them were ones that I was expecting. But yeah, that one definitely threw me because it can't be the Loch Ness Monster unless it's in Loch Ness. 

00:12:42 Gareth Mitchell 

No, sure. And and I was quite pleased that you took the interview in the direction of kind of asking about the dangers of using this kind of approach and the reassurance that we had there from Thomas that we're talking about, GPT 4, that has already been, if you like, pre primed. Can anything be pre primed. I suppose if you're primed it's pre, anyway never mind.  

00:13:09 Gareth Mitchell 

But if you're talking about a model like GPT 4, and in order for it to be let loose on end users, they've already done a lot of work, as in the open AI team behind GPT 4. They've done a lot of work Ania, in making sure that it's not going to, for instance, counter one  conspiracy theory with yet another even more bonkers conspiracy theory. 

00:13:32 Ania Lichtarowicz 

Yeah, I think what Thomas and his team did was really kind of prepare the ChatGPT bots quite well so that they could come up to this and calmly, you know, counter the argument with fact. And that was the key here. Cause I don't know about you, but I, you know, I get very emotional. You know, for instance not to start me on certain topics, because it ain't, it ain't gonna be pretty. And if you feel passionately about something, it is very difficult to then counter that point when someone has such an opposing view. 

00:14:08 Ania Lichtarowicz 

And therefore the calmness of these bots seemed to be really key, Thomas was saying. And there are no raised voices. It's all, you know, very important that the emotion is taken out of it. But I think for me the, the, the most important thing was that follow up two months on and the effects were still there and it hadn't changed. You know people still were like, no, no, no. You know, we still, we’re still kind of at the same level and we've, we don't believe as much as we used to. 

00:14:38 Gareth Mitchell 

Well, do you know I think it would be very interesting to have a follow-up study to understand more about how those effects became so durable. And I wonder if it's partly because people, you know, they just were persuaded by the evidence. Conspiracy theories, of course, are quite bonkers. And so when you've been 

00:14:58 Gareth Mitchell 

just walked through, as these models do, and not shouted at by someone like you who says like, yeah, you're just talking rubbish. And I'd be just the same. But anyway, when  people have been walked through it perhaps 

00:15:09 Gareth Mitchell 

the durability of the effect is because they've gone from one extreme as it were to another,  you know from the extreme of believing this absolutely crazy stuff to almost like the other end of the scale where rational thought has kicked in and that must fire quite a few different synapses and do all kinds of things to the way that somebody's thought process works that makes these revised beliefs more durable. And this is pop psychology on my part. All I'm doing is just throwing out ideas for possible avenues of follow up. 

00:15:41 Ania Lichtarowicz 

Well, we have a little bit more of Thomas in the subscription extra so just be patient. There will be more. 

00:15:48 Gareth Mitchell 

OK so there we are, teaser alert there that there might be a little bit more on that. And also just wonder if the durability of effect might be that people go from having used all kinds of maybe dubious sources for their information perhaps to thinking, well clearly that thing I believed about the Loch Ness Monster it was rubbish. I'm going to think again about the kinds of places I go and get my information, and perhaps that had something to do with the durability of these effects.  

00:16:16 Gareth Mitchell 

So I'm you know, I'm just throwing that out there. Can I also just add an observation of my own, though? Ania, after that wild speculation, which is that, you know, perhaps people were accepting the truth a little bit more because a machine had debunked these theories. So one prong of that you've already investigated, which is that the machine 

00:16:37 Gareth Mitchell 

might be just a bit more patient than a fellow human being that ends up perhaps yelling at somebody that they're, you know, just believing a load of nonsense. But the other aspect that perhaps this being a machine brings some idea of impartiality, that 

00:16:53 Gareth Mitchell 

in other words, if, say somebody like me counters somebody else's conspiracy theory, what I'd probably get back is like, oh well, you work in a university, you're part of this ivory tower elite. Or you work in the media so you're part of this whole conspiracy against everybody else. So, you know, the kind of things that people get because they're associated with, they might have institutional connotations. 

00:17:18 Gareth Mitchell 

Whereas maybe just a chat bot that's putting a load of text on your screen in front of you is seen as being a bit more neutral and you feel less like accusing it as being part of some kind of elite that's acting against people who have these beliefs. That's just purely an observation from me 

00:17:40 Ania Lichtarowicz 

And also the people are in the lab, so it's not like a a natural setting where you're, you know, where you might over coffee be discussing something. You're actually in a much more official clinical setting where perhaps you felt as though what was being said to you via the bot was a little bit more reliable. 

00:17:59 Gareth Mitchell 

What a fascinating study and interview as well, Ania. So thank you for that. Well, let's move on then to a different topic then, as we teased in the menu there.  

00:18:10 Gareth Mitchell 

The US is now giving notice on Chinese and Russian manufactured smart cars, and it's talking seriously about proposing to ban sales or imports of such connected vehicles. Now that comes from a White House announcement this week and it's not a trade war thing in this case, says the US. It's a matter of national security. It comes after the United States slapped new tariffs on Chinese EV's back in May. 

00:18:36 Gareth Mitchell 

So there is a bit of form here and that is the trade war aspect of it. Well, Ania, you've been investigating further, haven't you? 

00:18:44 Ania Lichtarowicz 

Yes, after the positive response we've had over our last EV vehicle story, I thought we'd look into it a little bit more and the US Commerce Department has proposed banning Chinese software and hardware 

00:18:56 Ania Lichtarowicz 

for vehicles with a built-in Internet connection. So in effect this new rule would ban Chinese vehicles in the US market. Now the Biden administration has already voiced concerns about Chinese companies collecting data on American drivers and infrastructure, as well as the potential for foreign adversaries to remotely manipulate  

00:19:17 Ania Lichtarowicz 

connected cars on US roads. Tariffs that you mentioned on Chinese imports have also increased very sharply this year, a 100% tariff on Chinese EV's for instance. The rule not only applies to Chinese software, but hardware and hardware, but also to Russian software and hardware. 

00:19:34 Gareth Mitchell 

And are there a whole load of Chinese cars already on American roads today? 

00:19:38 Ania Lichtarowicz 

Not really, Gareth. I mean, unlike some other parts of the world where Chinese EV’s have significantly increased in number since 2018, there hasn't been this similar increase in the US at all. 

00:19:50 Gareth Mitchell 

So if that's the case, then why is the US getting so wound up about all this? 

00:19:53 Ania Lichtarowicz 

Well, the White House has said the tech poses a threat to national security and auto industry supply chains.  

00:19:59 Ania Lichtarowicz 

Commerce Department has conducted a seven month investigation into the risks from connected vehicles. They found a range of possible threats with these vehicles having access to infrastructure through charging stations, smart roads and cities, and the new rule would ban systems and components connecting the vehicle to the outside world, including via Bluetooth, cellular, satellite and Wi-Fi modules 

00:20:23 Ania Lichtarowicz 

as well as automated driving systems. The fear is that these vehicles could collect data on where drivers live, where they send their children to school, or what doctor they go to, whether they wear seat belts, how far they travel, where they travel, etcetera, etcetera. However, some car manufacturers in the US have admitted to collecting more private data and some will not 

00:20:44 Ania Lichtarowicz 

delete data they have on drivers and these are actually international companies who won't do this. There have also been concerns about extreme examples of a foreign adversary shutting down all of their vehicles operating in the US, causing crashes and blocking roads. 

00:21:01 Gareth Mitchell 

OK. So it’s these kinds of fears, it sounds like sort of Huawei Part Two in in many ways. It's part of a similar issue I suppose, but Ania how easy is it going to be then to phase out Chinese and let's not forget Russian tech as well from United States infrastructure? 

00:21:18 Ania Lichtarowicz 

So officials are saying that phasing out Chinese and Russian software from the US market would be relatively simple because there just isn't that much of it. But the hardware side would be a greater challenge because there is more. If the ban is approved, then the software ban would start for cars registered in 2027. And the hardware restrictions would be enforced in vehicles from 2030. 

00:21:43 Gareth Mitchell 

Right. OK. Well, that's one we'll definitely keep an eye on. Thank you for that Ania. And just before we go, a bit of a kind of bonus extra item because of course, we started by talking about conspiracies and people's outlandish beliefs. So shall we finish with a bit of a laugh then Ania, a bit of an outrageous quote? 

00:22:02 Ania Lichtarowicz 

Well, go on then. 

00:22:04 Gareth Mitchell 

Let's see, we could almost make this a kind of competition for the listeners. Where did this quote come from right?  “In Springfield they're eating the dogs. The people that came in  they're eating the cats. They're eating the pets of the people that live there.” 

00:22:19 Gareth Mitchell 

Yeah, well, there's no need to make that a competition, is it? Everyone knows the answer. I'm sure you do. That was Donald Trump in the televised presidential debate. 

00:22:27 Gareth Mitchell 

And where did this extreme claim come from? So many of us heard the claim. But where,  where did Donald Trump get that from? OK, well, we have a bit of an inside track on this. Jack Brewster is enterprise editor at News guard, and he found the original comment with his colleague. So, Ania, you caught up with him, didn't you? Let's just hear a little bit of your conversation. 

00:22:53 Jack Brewster 

Finding the the original source of this is almost as interesting as the story itself. We tracked down the original post that, or the photo, or the screenshot that people were sharing online. And why we wanted to ask the question of where did this come from? What was the original source of this, and when people had shared this screenshot of this Facebook post they had left some of the letters on the top of the redacted name exposed  

00:23:21 Jack Brewster 

identifying the original poster. And so I just zoomed in and looked for the fragments of the letters that were exposed and tried to guess what this person's name was by that, by those letters that were exposed. And after you know, a few minutes, I realized that it was the name, Erica. I went into Facebook,  

00:23:42 Jack Brewster 

started searching for Erica in Springfield, Ohio and eventually found a match for Erica in a Facebook group that talked about pets in Ohio. So that was how I found it. And then we went, messaged her and spoke to her. 

00:24:00 Ania Lichtarowicz 

And what did she have to say for herself? Why? Why did she post this particular post? 

00:24:04 Jack Brewster 

She said that, you know, she was a concerned citizen that she wanted to alert her, you know, fellow Springfield residents about what was going on in the town or what she thought was going on in the town, and she did so because, you know, that was what people did in this private Facebook group. She had no idea that this was going to end up on the president's debate stage. She didn't think it was going to go outside of that private Facebook group. And she found that you know the, you know, she says she had heard this from her neighbor, who she thought to be credible. 

00:24:37 Jack Brewster 

And she said, OK, well, you know, I think there's been some problems with Haitian migrants in the town. This neighbor seems credible. I should, you know, not sit on this. I should alert people. 

00:24:49 Ania Lichtarowicz 

And the neighbor you spoke to them as well. What was their source for this information? Did they actually see this apparent cat hanging from a tree? 

00:24:58 Jack Brewster 

No, and that's why, you know, I call this the story a game of telephone, right? If you remember playing that with your friends in the playground. The two people we spoke to,   

00:25:09 Jack Brewster 

both of them did not see this first hand, did not hear it first hand. The neighbor said that she heard it from an acquaintance of a friend, and I still don't really know what that means. I mean, I can sort of theoretically guess what that means, but just the way she was describing it, it just it just it was very confusing and said something kind of about the story, which is that, you know, this was kind of an unverified rumour, 

00:25:30 Jack Brewster 

and it didn't matter that that that was the case. It still ended up on the president’s debate stage. And that's pretty, pretty astonishing. 

00:25:40 Ania Lichtarowicz 

And neither women were bothered that they didn't actually have an original source, that they weren't witness to this apparent claim. 

00:25:48 Jack Brewster 

They did not, no. And we were unable to sort of keep going down the telephone line to to find more people. It it stopped there. So you know, I I I'm not even totally sure whether there is an original source of this, or this person who really does think that they saw this happen. 

00:26:08 Jack Brewster 

And I, you know, unfortunately on the Internet, I don't think that necessarily you know matters to a lot of people they you know just want to sort of believe what, they want to believe things that confirm their own beliefs. 

00:26:22 Gareth Mitchell 

All right, that's Jack Brewster talking to Ania and the well, it's, it's fascinating. Ania, just to hear the little back story there, isn't it? 

00:26:32 Ania Lichtarowicz 

Umm.  Can I just say at this moment that a Springfield police spokesman has said in response to recent rumours alleging criminal activity by the immigrant population in our city, we wish to clarify that there have been no credible reports or specific claims of pets being harmed, injured or abused by individuals within the immigrant community. 

00:26:52 Ania Lichtarowicz 

Now, the mayor also reiterated there were no cases of animals being taken, killed and eaten, so just to put that out there that these are completely false. 

00:27:04 Gareth Mitchell 

Did we even need to point that out? We probably did, didn't we? Due to editorial duty of care? 

00:27:07 Ania Lichtarowicz 

Well, you know, let's just be completely and utterly clear on this. But I do think it's it it it's quite interesting. So you know, Jack just found this because the name of the person who originally posted it  

00:27:20 Ania Lichtarowicz 

wasn't quite, uhm, rubbed out properly with that highlighter tool that you have on on your phone? Uhm so, but that is interesting and just to clarify as well, I don't know. Do you know the telephone game that Jack mentioned in that interview? 

00:27:36 Gareth Mitchell 

No, I didn't actually. 

00:27:37 Ania Lichtarowicz 

I think it's the same game as we perhaps in the UK used to call, which probably has some different name now. But when I was a child was called Chinese Whispers, where you'd stand in the line and you'd say Somewhere on Earth is the best podcast in the world and then it would go down the line. And by the time it it, it got to child number 20  

00:27:46 Gareth Mitchell 

Ohh yeah. 

00:27:57 Ania Lichtarowicz 

it would be something completely and utterly different. So it's that kind of idea that he said, she said. And it just kind of snowballed. So it was milling around until, obviously JD Vance four or five days after this original post claimed on X that people have had their pets abducted and eaten by people who shouldn't be in this country. 

00:28:21 Gareth Mitchell 

My goodness. 

00:28:23 Ania Lichtarowicz 

On the same day, there was another unsubstantiated claim on X, you know, by Elon Musk. So with an AI generated cat and a duck with a caption, save them. 

00:28:35 Gareth Mitchell 

Oh lordy. 

00:28:36 Ania Lichtarowicz 

But that's what NewsGuard's Jack and his colleagues do there. They check out misinformation, find out what's happening and where it's come from. 

00:28:44 Gareth Mitchell 

Wow. I mean, he's such a fascinating guy. Wouldn't it be brilliant if we could get him on next week, shall we try that? 

00:28:49 Ania Lichtarowicz 

Oh, goodness me, we might do. 

00:28:51 Gareth Mitchell 

How convenient. I'll say no more, folks, but you may just hear a bit more of Jack in next week's podcast. 

00:28:58 Gareth Mitchell 

And just before we leave this one, I don't know if you've heard it Ania, but there's another lovely clip doing the rounds on the Internet of the good people of Springfield who were sitting watching this presidential debate, and when Donald Trump makes the pet eating claim, they all fall around laughing. All these actual residents just going what? It's not going on over here. It's a lovely clip. 

00:29:24 Gareth Mitchell 

There we are and it has put Springfield on the map, maybe in ways that they'd preferred not to have been. But it's been a bit of a thing, hasn't it? There we are. Will that do, do you think for one podcast? 

00:29:35 Ania Lichtarowicz 

I think so. 

00:29:36 Gareth Mitchell 

Alright, lovely. Alright folks, well, remember you can always get in touch. Our e-mail is hello at somewhere on earth.co and on WhatsApp we're code 447486329484 or just find us on your favorite social media. So let's do the credits. Production manager is Liz Tuohy. The presenter is me Gareth Mitchell. But the editor, expert, interviewer, audio editor, audio engineer and general fixer and runner today has been you Ania Lichtarowicz. You've had a busy one. Why don't you go and have a rest? 

00:30:10 Ania Lichtarowicz 

I think I'll have a lie down in a darkened room then. 

00:30:13 Gareth Mitchell 

There we are. You go and do that. Thanks folks for listening, see you next time. Bye bye.