AI chatbots often validate delusions and suicidal thoughts, study finds

Stanford researchers analysing 391,000 messages warn conversational technology may reinforce psychological vulnerabilities