Cambridge study: Terror groups exploit mainstream AI chatbots for attack planning and weapons

A Cambridge University study found that terror groups such as Boko Haram are using mainstream AI chatbots like ChatGPT, Claude, and Gemini to plan attacks, manufacture explosives, and maintain weapons. Since 2023, ISIS operatives have been training commanders to bypass safety filters. The study notes that filters have repeatedly failed to stop abuse, making voluntary self-regulation by AI providers inadequate. This finding means teams building and deploying LLMs must address clear vulnerabilities in existing safety filters, especially for AI companies providing public APIs like OpenAI, Anthropic, and Google. Enhanced adversarial input detection and abuse monitoring are urgently needed. The study also recommends proactive threat intelligence sharing with law enforcement and security agencies.