AI Induced Psychosis: A shallow investigation — LessWrong

In this post, I took a brief dive into how different AIs respond to a user who has increasingly severe psychosis symptoms. I rated different AIs based on how often they push back against users, how often they confirm the user’s delusions, and how much they conform to therapeutic best practices in dealing with psychosis patients. I also read through red teaming transcripts and extracted some illuminating quotes on how different AIs respond. I think incorporating multi-turn AI red teaming pipelines and professional therapeutic guidance into the post training process could do a lot to reduce AI-driven psychosis.
— Read on www.lesswrong.com/posts/iGF7YcnQkEbwvYLPA/ai-induced-psychosis-a-shallow-investigation