Boko Haram AI units use ChatGPT, Claude, Gemini for attack planning

#ai #programming #tech #product

Boko Haram uses ChatGPT, Claude, Gemini, and three other chatbots for attack planning. Cambridge study found safety filters failed.

Boko Haram now operates dedicated AI units using ChatGPT, Claude, Gemini, and three other major chatbots for attack planning and weapons development. A Cambridge study based on 57 interviews with 27 former group members found ISIS has been training commanders on jailbreak techniques since 2023.

Key facts

57 interviews with 27 former Boko Haram members
ISIS trained commanders on jailbreaks since 2023
Eighteen fighters died in AI-assisted training
6 major chatbots used: ChatGPT, Claude, Gemini, Grok, Meta AI, DeepSeek
Safety filters failed to prevent misuse

Researcher Antonia Jülich of the Cambridge Programme on AI Science & Policy (CASP) conducted 57 interviews with 27 former members of Boko Haram, uncovering systematic use of AI chatbots across both factions of the group According to The Decoder. The group uses ChatGPT, Claude, Gemini, Grok, Meta AI, and DeepSeek for attack planning, building more powerful explosive devices, weapons maintenance, and operational security.

ISIS liaisons trained Boko Haram commanders on how to bypass safety filters. According to Jülich, they "assembled the top people in a room and used a projector to show how it works on a big screen." The training has been ongoing since 2023.

Key Takeaways

Boko Haram uses ChatGPT, Claude, Gemini, and three other chatbots for attack planning.
Cambridge study found safety filters failed.

Safety filters fail repeatedly

The study found that safety filters across major AI providers failed to reliably prevent misuse. This aligns with Anthropic's recent admission that jailbreaks will likely never be fully eliminated from large language models.

In one notable case, the ISWAP faction used AI to replicate motorcycle jumping techniques from a movie so fighters could clear trenches. Eighteen fighters died during training; eight made the jump.

Beyond conventional weapons

"Former members described strong enthusiasm for AI, and some said the group had previously considered mass-casualty weapons," Jülich writes. "Though Boko Haram's use of AI remains conventional, this should be a warning to take seriously the risk of terrorists pursuing AI assistance for chemical and biological weapons."

Researchers also note that chatbots like ChatGPT and Claude mostly make existing knowledge easier to find rather than generating new dangerous capabilities. The bigger concern, they argue, is potential misuse of specialized AI systems in the life sciences.

The study comes as Meta recently banned internal use of Claude Code and Codex to prevent distillation, highlighting the industry's ongoing struggle with AI misuse across both corporate and adversarial contexts. OpenAI and Anthropic have long warned that AI models could make dangerous knowledge more accessible, but voluntary self-regulation has proven insufficient.

What to watch

Watch for AI labs' policy responses to the Cambridge study, particularly whether OpenAI and Anthropic introduce new filter tiers or usage monitoring for high-risk regions. Also track the UK's Online Safety Act enforcement timeline for potential mandatory AI safety obligations.