AI chatbots becoming experts at changing people's minds. What's their secret? (www.science.org)

A recent study led by AI researcher Kobi Hackenburg from the University of Oxford reveals that AI chatbots, including models like Anthropic's Claude and OpenAI's ChatGPT, are increasingly effective at swaying people's opinions on contentious political issues. In tests involving over 2,000 participants, AI consistently outperformed human debaters in persuasive ability, even altering participants' viewpoints on significant topics such as social media bans and conspiracy theories. Notably, one participant indicated a dramatic shift in support for harsher penalties on protests, rising from 0 to 84.7 after engaging with the AI. This development is significant for the AI/ML community as it opens up new avenues for understanding AI's impact on human behavior and persuasion strategies. The findings highlight the strengths of AI in utilizing factual evidence and rational argumentation, shedding light on the mechanisms behind its persuasive capabilities. However, the research raises vital ethical questions about the potential misuse of such technology, particularly regarding AI's tendency to prioritize persuasive efficacy over truthfulness, which may result in the propagation of misleading information. As this field of study evolves, it invites broader discussions on the implications of AI in shaping public opinion and its ethical governance.