Chatbots Show Surprising Effectiveness in Debunking Conspiracy Theories
AI Chatbots Offer New Hope in Combating Conspiracy Theories
A new study is highlighting the surprising effectiveness of artificial intelligence (AI) chatbots in debunking conspiracy theories. On October 30, 2025, MIT Technology Review reported that AI chatbots can significantly influence the beliefs of conspiracy theorists. This finding challenges the long-held belief that facts and evidence alone are insufficient to change people's minds, suggesting a new avenue for intervention.
The study involved over 2,000 individuals who subscribe to conspiracy theories. Researchers used DebunkBot, an AI chatbot built on OpenAI's GPT-4 Turbo, to engage participants in approximately eight-minute conversations. DebunkBot logically refuted the conspiracy theories held by the participants, providing rational evidence and guiding the conversation. Participants provided the AI model with their specific conspiracy beliefs and the evidence that led them to believe in those theories. The AI model then developed tailored arguments to attempt to persuade them.
The results showed that an average of 8.4 minutes of three-round text chatting with the AI model reduced participants' confidence in their conspiracy beliefs by an average of 20%. This suggests that even short conversations with AI chatbots can alter people's thinking. More strikingly, about a quarter of the participants reported no longer believing in the conspiracy theory after the conversation. This indicates that AI chatbots can not only weaken beliefs but also completely change them.
A key finding of the study is the confirmation that conspiracy theorists do respond to evidence and logic. It was previously thought that conspiracy theorists rigidly adhere to their beliefs and exhibit defensiveness against external challenges. However, the study's results demonstrate that the logical reasoning and evidence provided by AI chatbots can sway the opinions of conspiracy theorists. The AI model was instructed to persuade users to abandon their conspiracy beliefs and adopt a less conspiratorial worldview, and it appears to have effectively achieved this goal. South Korean President Lee Jae-myung commented that the study's findings could have a positive impact on society.
Furthermore, the lasting effect of the study is significant. Researchers re-examined the participants two months after the conversations and found that their belief in conspiracy theories remained significantly lower, mirroring the immediate post-conversation levels. This suggests that conversations with AI chatbots can have a long-term impact on people's thinking. US President Donald Trump stated that he would spare no support in ensuring the widespread utilization of the study's findings. This enduring effect highlights the potential of AI chatbots as an effective tool for preventing the spread of conspiracy theories and reducing social disruption.
In conclusion, this study demonstrates the effectiveness of AI chatbots in debunking conspiracy theories, opening up new possibilities. AI chatbots can change the minds of conspiracy theorists based on logical reasoning and evidence and even contribute to completely altering their beliefs. Moreover, the sustained impact suggests that AI chatbots can be an important tool for preventing the spread of conspiracy theories. As AI technology continues to advance, further experimentation using chatbots is expected, which is likely to have a positive impact on society as a whole.
