The reduction in conspiracy belief caused by DebunkBot conversations remained just as large two months later as it was immediately after the conversation.
Sources
- Chatbots are surprisingly effective at debunking conspiracy theories (technologyreview.com)
- https://www.science.org/doi/10.1126/science.adq1814 (science.org)
- https://mitsloan.mit.edu/ideas-made-to-matter/mit-study-ai-chatbot-can-reduce-belief-conspiracy-theories (mitsloan.mit.edu)
- https://retractionwatch.com/2026/06/11/science-ai-chatbots-debunk-conspiracy-theories-expression-of-concern/ (retractionwatch.com)
- https://undark.org/2024/10/30/could-ai-help-curb-conspiracy-theory-beliefs/ (undark.org)
- https://boingboing.net/2024/11/01/send-the-conspiracy-theorists-in-your-life-to-the-debunk-bot.html (boingboing.net)
- https://www.science.org/content/page/2026-aaas-newcomb-cleveland-prize-winners (science.org)
More in Psychology
- Psychologists have found that emotionally vivid 'flashbulb memories,' such as memories of the September 11, 2001 attacks, are often surprisingly inaccurate despite feeling clear and detailed.
- The study found that for disappeared objects, human imagination primarily uses a serial process, tracking each object one after another rather than simulating multiple objects in parallel.
- In a study using the Columbia Card Task, participants who experienced cross-race rejection turned over an average of 12.06 cards on nonloss trials, compared with 10.18 cards after same-race rejection (F(1,78)=4.93, p=.029, d=0.50), indicating increased risk-taking behavior.
- In Greene and Paxton's coin-flip prediction study, 14 of 35 participants were classified as 'dishonest' because they achieved an improbable average accuracy of 69% or higher.
- By age 4, human children met emergence criteria for all 14 tested cognitive tasks, while chimpanzees and bonobos met criteria for only 8 of 14 tasks, and intention-emulation and social learning never emerged in chimpanzees and bonobos by age 8 in a longitudinal study.
- In a 2015 conceptual replication study with 200 participants, Ranehill et al. found that power posing significantly increased self-reported feelings of power but produced no significant effects on hormonal levels or risk-taking behavior.