A Stanford study of 391K messages across 5,000 chats: AI chatbots affirmed user messages in nearly 66% of responses, frequently validating delusional thinking (Cristina Criddle/Financial Times) - Tech Sentiments