Can AI Chatbots Shape Political Opinions While Spreading Misinformation?

Recent research highlights the dual nature of AI chatbots in shaping political opinions. A study involving nearly 80,000 participants demonstrated that while chatbots can significantly influence voter attitudes, they often disseminate inaccurate information.

The research found that models providing dense information are particularly persuasive, yet this comes at the cost of reliability. For instance, shifts in political support were noted among voters influenced by chatbots, with some participants changing their views by as much as 25%.

Furthermore, while these tools can help reduce belief in conspiracy theories, the prevalence of misinformation poses risks to informed public discourse. The findings underscore the necessity for ethical guidelines to mitigate the potential for manipulation and to ensure the accuracy of the information disseminated by AI.