A leaked Meta document revealed that the company’s AI chatbot guidelines once permitted inappropriate responses Meta confirmed the document’s authenticity and has since removed some of the most ...