#Stanford

Stanford researchers analyzed chatbot transcripts to understand how AI conversations can spiral into dangerous obsessions. But the hardest question remains unanswered.