14 September

Very interesting research shows that if conspiracy theorists engage in long conversations with AI bots, their beliefs can be shifted in ways that doesn’t happen if they’re only presented with alternative facts, or have long arguments with humans. The researchers say they think this is because conspiracy theorists are just trying to make sense of things, like everybody else. But I’m sure that’s not the reason. No amount of talking to humans budges them. I think the AI bots succeed because they’re not a threat, socially. They’re not part of the opposing ‘team’.