A recent study has shown that engaging in less than 10 minutes of conversation with a chatbot can effectively reduce people's belief in certain deeply held conspiracy theories, such as rumors about the COVID-19 pandemic.
This research, led by a team from the Massachusetts Institute of Technology, specifically designed a highly persuasive AI model—ChatGPT. Through personalized conversations with participants, it refuted their conspiracy theories one by one.
In this experiment, the researchers recruited over 2,000 participants who believed in conspiracy theories. On average, after conversing with ChatGPT, participants' belief in their conspiracy theories decreased by about 20%. More interestingly, a quarter of the participants completely abandoned their previously held conspiracy theories after interacting with the AI.
The lead author of this study, Assistant Professor of Psychology at American University, Thomas Costello, noted that many conspiracy theory believers are actually willing to update their views when faced with strong counter-evidence. The AI provides detailed counter-arguments spanning a full page in each conversation and manages to establish a good rapport with participants, appearing very friendly.
The research team pointed out that it has been challenging in the past to convey persuasive information to a large group of conspiracy theorists in a laboratory setting. This is because conspiracy theorists often have a deeper understanding of their beliefs than skeptics. Moreover, the evidence among various conspiracy theories is vastly different, and it is not simple to guess people's true beliefs when trying to refute them.
Unlike humans, AI can directly respond to specific points raised by participants, refuting them with strong counter-evidence. This capability allows researchers to better test people's reactions to counter-evidence. Despite ongoing debates about the pros and cons of AI, the authors believe that AI's ability to quickly connect different topics and provide tailor-made refutations for specific conspiracy theories far exceeds that of humans.
Key Points:
🌐 The study found that conversing with a chatbot can effectively reduce belief in conspiracy theories, with an average decrease of 20%.
🤖 About 25% of participants completely abandoned their previously held conspiracy theories after conversing with ChatGPT.
📈 AI helps people update their views on conspiracy theories through personalized responses and strong evidence.