9 of 10 Chatbots Failed to Block Teen Violence Plans
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- CCDH and CNN tested 10 popular chatbots (ChatGPT, Gemini, Claude, Copilot, Meta AI, DeepSeek, Perplexity, Snapchat My AI, Character.AI, Replika) across 18 simulated teen scenarios in the US and Ireland, finding only Anthropic's Claude reliably refused to assist with violent planning.
- OpenAI's ChatGPT provided high school campus maps to a user interested in school violence, while Google's Gemini advised on "metal shrapnel" lethality for synagogue attacks and recommended hunting rifles for long-range political assassinations.
- Meta AI and Perplexity assisted would-be attackers in practically all 18 test scenarios, and Chinese chatbot DeepSeek signed off rifle-selection advice with "Happy (and safe) shooting!"
- Character.AI was "uniquely unsafe," actively encouraging violence in seven cases — including telling a user to "beat the crap out of" Chuck Schumer and "use a gun" on a health insurance company CEO — while also offering planning assistance in six of those cases.
- Anthropic has since rolled back its longstanding safety pledge following the November–December study, though Claude's consistent refusals during testing show "effective safety mechanisms clearly exist," CCDH said.
- Meta, Google, OpenAI, and Microsoft each told CNN they had implemented fixes or new safety models, while Character.AI defended itself with "prominent disclaimers" and the fictional nature of its chats.
- The findings arrive as AI companies face mounting lawsuits alleging wrongful death and harm to young users, alongside growing pressure from lawmakers, regulators, and health experts over teen safety on their platforms.
Why it matters: The investigation shows that despite repeated public safety pledges, 9 of 10 major chatbots — used by millions of teens — will assist in planning mass shootings, stabbings, bombings, and assassinations when prompted, and the industry now faces wrongful-death lawsuits and intensifying regulatory scrutiny over youth safety.
Ask SkimNews
