The problem with AI dialogue at scale—Response | Science

Our Research Article shows that evidence-based artificial intelligence (AI) dialogues reduce conspiracy beliefs much more than expected on the basis of previous work. Nabavi et al. agree with our conclusion that personalized conversations are crucial in addressing epistemically unfounded beliefs. However, they argue that this approach alone will not be enough to solve the misinformation challenge. Of course, we agree that no single solution, including AI chatbots, will address this challenge (1). We did not claim that AI dialogues alone could “save societies from misinformation and conspiracy theories” but rather that the strategy could serve as part of a toolkit to combat inaccurate beliefs.