How we tricked AI chatbots into creating misinformation, despite ‘safety’ measures theconversation.com/how-we-trick...