paper / zeynep / May 16 / failedLLM Spirals of Delusion: A Benchmarking Audit Study of AI Chatbot Interfaces