Authors' Introduction
- Conspiracy theories are seemingly everywhere. To explain their prevalence, many commentators point to the gullibility of conspiracy theorists. According to this view, believers in conspiracy theories accept evidence without bothering to scrutinise its credibility, making them vulnerable to the misinformation that pervades online ecosystems. But while it’s tempting to take this view, we believe it relies on an unrealistic picture of misinformation and the people who consume it – which is likely undermining attempts to deal with the problem.
- Far from passively accepting the truth of conspiracy theories, conspiracy theorists enthusiastically participate in generating, discussing and dissecting them. They also appear genuine in their attempts to get to the bottom of things. They develop sophisticated arguments, go to considerable lengths to find the ‘right’ sources of information, and preach the importance of rigorous and independent research. Conspiracy theorists don’t fall for conspiracy theories. They discover them.
Authors' Conclusion
- One approach to delivering insight-based interventions could be via developments in AI, which hold the promise of countering conspiracy thinking via personalised interactions. For example, in a recent study, interactions with ChatGPT-4 Turbo substantially and durably shifted the beliefs of even the staunchest of believers in conspiracies ranging from the causes of COVID-19 to the Moon landings to the death of Princess Diana. Analysing these conversations showed that the model didn’t simply present counterevidence but questioned and reasoned with users. Although users were not explicitly asked to report insight experiences, we would venture that such experiences were a crucial mechanism underlying their shift in beliefs.
- The mind-changing potential of these systems is bound to continue to grow. They already have more ‘patience’ than any human for gaining our idiosyncratic perspective through dialogue. As they can ingest more information about us, they will be able to simulate our positions more effectively. That will allow them to challenge us at the just-right level – eventually coming up with new, resonating metaphors and pointed questions. We are hopeful that such personal(ised) AI assistants will help to prevent and counter harmful beliefs at scale (but they will need to be implemented carefully, given the risk of misuse).
- Conspiracy believers are not unintelligent and gullible. They are driven by their hunger for insight. Whether via AI or other media, recognising and respecting this hunger is the way towards more effective interventions.
Author Narrative
- Stephen Gadsby is a Fonds Wetenschappelijk Onderzoek (FWO) postdoctoral fellow at the Centre for Philosophical Psychology at the University of Antwerp in Belgium.
- Sander Van de Cruys is a researcher in psychology based in Belgium. In his work, he tries to understand the internal rationality of supposedly irrational thinking and behaviour.
Notes
- I think the authors are too kind to conspiracy theorists.
- I think there are two sorts of people who fall into this category: leaders and followers.
- Followers are - I suspect - often uncritical, lazy thinkers who support ideas that go along with their world view or social needs.
- Leaders may well be highly intelligent and innovative, but they suffer from the same defect as followers, insofar as they are reading the world so it aligns with their prejudices.
- I suspect both camps are lazy in another respect. Being an expert in any technical discipline takes years of training in addition to intelligence and innovation. Conspiracy theorists may take the easy way out and start from scratch. They ignore the experts who have earned the right to have an informed opinion.
- Contrast with "Cassam (Quassim) - Bad thinkers" which is referenced (probably by the editor) and which takes a more negative stance.
- There are no Aeon Comments, but a few more links to be followed up.
Comment:
Text Colour Conventions (see disclaimer)
- Blue: Text by me; © Theo Todman, 2026
- Mauve: Text by correspondent(s) or other author(s); © the author(s)