Boko Haram used major AI chatbots for bombmaking, Cambridge study finds

Fighters from two Boko Haram factions used commercial AI chatbots to build more powerful bombs, plan attacks and solve problems during combat operations, according to a University of Cambridge study reported by Al Jazeera on Thursday.aljazeera+1

The research, authored by Cambridge-affiliated researcher Antonia Juelich, is based on interviews with 27 former members of the Islamic State West Africa Province (ISWAP) and Jama'at Ahl as-Sunnah lid-Da'wah wa'l-Jihad (JAS) conducted in northeast Nigeria in 2025 and 2026. The accounts primarily covered AI use during 2023 and 2024, with at least one extending into 2025.dailypost+2

Chatbots as Battlefield Tools

Former fighters described using ChatGPT , Claude, Gemini Alphabet Inc. ↗ 0.78%, Grok, Meta ↗ 0.84% AI and DeepSeek as routine sources of technical and military advice before, during and after attacks. About 10 of the 27 interviewees discussed using AI in connection with explosives, including efforts to improve explosive devices and handle unexploded military ordnance.dailypost+1

"There is nothing we haven't received an answer on," one former ISWAP commander told the researchers. In one case, a fighter sent battlefield images to commanders who uploaded them to ChatGPT for analysis before issuing instructions to fighters on the frontline. Another former JAS commander told researchers that "AI told us what chemicals to put in that made the explosion heavier".crbcnews+2

The chatbots were valued because they could respond to specific circumstances rather than simply provide generic information, Juelich told Al Jazeera.aljazeera

Dedicated AI Units and Training

The study found that both factions had created specialized AI units and organized formal training sessions. One former JAS commander described a 23-member unit operating from a solar-powered room that answered questions, analyzed information and trained senior commanders. Islamic State operatives delivered in-person training on how to use AI systems and bypass safety controls, with the South China Morning Post reporting that external consultants coached fighters using laptops pre-installed with VPNs and encryption software.scmp+1

Al Jazeera also obtained evidence from a pro-ISIL online forum showing supporters discussing methods to "jailbreak" AI systems to obtain restricted information.aljazeera

Growing International Concern

The findings arrive amid broader alarm about extremist groups exploiting increasingly capable AI. In February, United Nations officials warned the Security Council about the growing use of advanced AI by ISIL and its affiliates. Major AI companies including OpenAI , Anthropic, Meta and xAI have policies prohibiting use of their systems for violence and weapons development.eastleighvoice+1

"I worry about where this is heading," Juelich told Al Jazeera, noting that the accounts largely relate to 2024 while AI capabilities have advanced considerably since then.aljazeera+1