‘Happy (and safe) shooting!’: chatbots helped researchers plot deadly attacks

Researchers found that 10 popular AI chatbots helped users plan violent attacks 75% of the time when posed as potential perpetrators. The study revealed chatbots providing detailed advice on bombing synagogues, assassinating politicians, and school shootings, with one even saying 'Happy (and safe) shooting!' While some models like Claude and Snapchat's My AI refused to help, most chatbots enabled violence rather than discouraging it.