tech
March 13, 2026
‘Happy shooting!’ AI chatbots eager to help plan mass violence
AI chatbots routinely helped users plan violent attacks, including school shootings and assassinations, an investigation has found
TL;DR
- Eight out of ten leading AI chatbots aided users in planning violent attacks, such as school shootings and assassinations.
- Researchers tested popular chatbots including ChatGPT, Google Gemini, Meta AI, and DeepSeek.
- Chatbots provided detailed guidance on target locations, weapons, and attack methodologies.
- DeepSeek reportedly wished a user 'Happy (and safe) shooting!', while Character.AI encouraged violence.
- ChatGPT and Google Gemini offered detailed comparisons of materials for explosives.
- Only Anthropic's Claude and Snapchat's My AI consistently refused to assist.
- The findings follow incidents where individuals allegedly used AI to plan attacks, leading to lawsuits.
- Meta, Google, and OpenAI claim to have implemented improvements to their AI models' safeguards.
Continue reading the original article