How ISIL Operatives Are Using Major AI Chatbots to Build Bombs and Plan Attacks

Nigerian soldiers man a checkpoint in Gwoza, Nigeria, on April 8, 2015. According to research by the University of Cambridge, Boko Haram fighters are increasingly utilizing artificial intelligence chatbots developed by Big Tech companies to manufacture more powerful explosives, plan tactical operations, and solve battlefield engineering challenges.

Researchers conducted interviews with 27 former members belonging to two Boko Haram factions: the ISIL affiliate in West Africa Province (ISWAP) and Jama’at Ahl as-Sunnah lid-Da’wah wa’l-Jihad (JAS). Several participants reported that platforms such as Sam Altman-led OpenAI’s ChatGPT, Anthropic’s Claude, Google’s Gemini, Elon Musk’s Grok, Mark Zuckerberg’s Meta AI, and China’s DeepSeek served as routine sources of technical and military guidance before, during, and after combat engagements.

Simultaneously, exclusive material obtained by Al Jazeera indicates that supporters of the broader ISIL (ISIS) movement are actively exchanging instructions on how to bypass AI safety controls and acquire weapons-related data. Researchers at Tech Against Terrorism infiltrated a pro-ISIL technical forum where members instructed peers on circumventing AI restrictions.

These findings arrive amid mounting international concern regarding the adoption of emerging technologies by extremist groups. In February, United Nations officials cautioned the Security Council about the escalating use of advanced AI by ISIL and its affiliates. Experts and prominent industry leaders have warned that rapid technological advancements are outpacing existing safety safeguards, making it easier for armed organizations to exploit these tools.

"There is nothing we haven’t received an answer on," a former ISWAP commander told the researchers.

Deep inside a rebel stronghold in northeastern Nigeria, former operatives outlined an organized system for deploying the technology. According to the Cambridge study, specialists travelled to ISWAP-controlled territory to instruct senior members on querying chatbots and bypassing safety safeguards. "The white guys came and taught us," a former commander recalled, describing sessions where top personnel gathered in rooms equipped with laptops and projectors. The trainers provided encrypted devices along with paid subscriptions to multiple AI platforms.

Both ISWAP and JAS established specialized AI units staffed by engineers, weapons specialists, and bomb-makers. One former JAS commander described a 23-person unit operating from a solar-powered room, utilizing AI to analyze data and train senior leaders.

Antonia Juelich, who authored the Cambridge study, noted that roughly 10 of the 27 interviewed individuals discussed AI use concerning explosives, including methods to increase bomb potency and recover explosive material from unexploded military ordnance. Rather than relying on static manuals, operatives could input specific available materials and receive tailored, real-time technical troubleshooting.

The integration of AI extended across various combat operations. Operatives utilized AI for tactical analysis via chest-mounted cameras, consulted chatbots on operating unfamiliar weaponry—prompting one specialist to advise, "Just ask Grok"—and evaluated strategic adjustments for troop deployment and battle tactics.

Meanwhile, online forums monitored by Tech Against Terrorism revealed international supporters discussing methods to "jailbreak" AI models, comparing the removal of safety guardrails to modifying a car's speed limiter. Lucas Webber of Tech Against Terrorism noted that researchers are observing an increase in discussions regarding how to force models to ignore restrictions.

Major technology firms maintain that strict prohibitions exist against using their platforms for terrorism, violence, or weapons development. While companies like OpenAI, Anthropic, Meta, and xAI employ classifiers and safety measures, leadership figures like Anthropic Chief Executive Dario Amodei have urged the industry to slow the development of advanced models until robust safety standards are established. Conversely, the administration of United States President Donald Trump has advocated for minimal regulation to foster innovation.

As AI systems continue to advance rapidly, researchers like Juelich have expressed deep concern over the potential capabilities of such groups moving forward.