Please join us for the inaugural LAMBDA (Looking After Minds and Brains in the Digital Age) Dialogues session focusing on the psychological and sociocultural impacts of AI (chaired by Dr Tom Pollak and Dr Hamilton Morrin).

Jared Moore, a PhD student in Computer Science at Stanford University, will join us remotely to present his research examining how large language models (LLMs) can amplify delusional spirals and the potential implications of AI for mental health and wellbeing.

Abstract: As large language models (LLMs) have proliferated, disturbing anecdotal reports of negative psychological effects, such as delusions, self-harm, and “AI psychosis,” have emerged in global media and legal discourse. In our work, we analyze logs of conversations with LLM chatbots from 19 users who report having experienced psychological harms from chatbot use. We develop an inventory of 28 codes and apply it to the 391,562 messages in the logs. We find, for example, that messages that declare romantic interest and messages where the chatbot describes itself as sentient occur much more often in longer conversations, suggesting that these topics could promote or result from user over-engagement and that safeguards in these areas may degrade in multi-turn settings. We also built an evaluation protocol, DelusionEval, to assess how LLMs respond to messages with prior context from these real transcripts. In this evaluation, we show that, while some delusion-linked behaviors have decreased with larger and more recent LLMs, rates remain high.

Registration: from 17:00Lecture: 17:30–18:30Q&A: 18:30–19:00, under the Chatham House RuleClose: 19:00

All are welcome. This lecture will be of particular interest to clinicians, researchers, developers and others interested in the intersection of artificial intelligence and mental health.