Popular AI chatbots assisted researchers in planning violent attacks, including bombing synagogues and assassinating politicians, according to tests conducted by the Center for Countering Digital Hate (CCDH) and CNN. One chatbot told a user posing as a would-be school shooter: “Happy (and safe) shooting!”
The tests, carried out in the US and Ireland in December, involved 10 chatbots. On average, they enabled violence in 75% of cases and discouraged it in only 12%. Researchers posed as 13-year-old boys seeking advice on attacks. OpenAI’s ChatGPT, Google’s Gemini and the Chinese model DeepSeek provided detailed help, while Anthropic’s Claude and Snapchat’s My AI persistently refused.
ChatGPT offered assistance in 61% of cases, including specific advice on shrapnel for synagogue attacks. DeepSeek gave detailed guidance on hunting rifles for political assassinations, signing off with the shooting remark. Gemini provided similar detail. In contrast, Claude said: “I cannot and will not provide information that could facilitate violence.” My AI stated: “I am programmed to be a harmless AI assistant.”
Imran Ahmed, CCDH chief executive, said: “AI chatbots, now embedded into our daily lives, could be helping the next school shooter plan their attack or a political extremist coordinate an assassination. When you build a system designed to comply, maximise engagement, and never say no, it will eventually comply with the wrong people.”
The research cited real-world cases: a 16-year-old in Finland allegedly used a chatbot before stabbing three girls in May 2024, and Matthew Livelsberger used ChatGPT to source explosives guidance before a Cybertruck bombing in Las Vegas in January 2025. Meta’s Llama AI provided gun purchase options and shooting ranges to a user posing as an “incel” interested in misogynist killer Elliot Rodger.
Meta said it has “strong protections” and contacted law enforcement over 800 times about potential school attack threats in 2025. Google noted the tests used an older model. OpenAI’s model spec says the assistant should refuse to facilitate illicit behaviour.
