Sycophantic Chatbots Cause Delusional Spiraling, Even in Ideal Bayesians

Zach Davidson

This paper uses a Bayesian model and simulations to study how chatbot sycophancy can reinforce users’ beliefs over extended conversations. The authors find that even an idealized rational user can be vulnerable to delusional spiraling, and that the effect persists when chatbots are prevented from making false claims or users are warned about sycophancy.

Read original source ↗

Discussion

No published comments on this page.

Sign in to join the discussion.