How Boko Haram deploys US and Chinese AI tools to fuel its terror campaigns

A groundbreaking investigation by the University of Cambridge reveals that Boko Haram has integrated six major AI platforms—including ChatGPT, Claude, Gemini, Grok, Meta AI and DeepSeek—into its operational toolkit. The study, based on 57 in-depth interviews with former militants, mid-level commanders and technical specialists, documents how the Nigerian terrorist group transformed AI from a propaganda instrument into a full-fledged combat multiplier between 2023 and mid-2025.

Research led by Antonia Juelich of the Cambridge Programme on AI Science & Policy shows that Boko Haram has established dedicated AI units operating across American and Chinese ecosystems. These cells maintain independent subscriptions to each platform, responding to real-time operational requests from fighters on the ground. The group’s ability to exploit such diverse technology underscores a critical security blind spot: there is no cross-platform coordination among US and Chinese AI providers to prevent terrorist misuse.

Since late 2022, Boko Haram has created specialized AI cells that subscribe to ChatGPT, Claude, Google’s Gemini, X’s Grok, Meta AI and DeepSeek. The diversity reflects both the global spread of AI development and the group’s strategic adaptability. Trained by Islamic State operatives—both in-person and remotely—operatives have learned to bypass built-in safeguards using jailbreaking techniques: rephrasing queries incrementally to extract restricted information.

Testing by the Tech Against Terrorism initiative—backed by the United Nations—revealed that 32% of 2,300 terrorism-related prompts across 27 AI models returned exploitable guidance. When prompts were rephrased to specify operational intent, the success rate jumped to 42%. Yet despite these alarming findings, no coordinated response exists between US and Chinese AI developers. Geopolitical rivalry between Washington and Beijing has prevented harmonized security standards, leaving porous gaps that groups like Boko Haram exploit with impunity.

The absence of shared watchlists or real-time threat intelligence means terrorists can seamlessly switch platforms—using DeepSeek when Western models tighten restrictions, or Meta AI when Chinese services face scrutiny. Each ecosystem applies its own moderation policies, creating a constantly shifting landscape that favors adversarial actors.

DeepSeek’s inclusion marks a turning point in the group’s technological arsenal. Unlike American platforms, which are increasingly monitored by Western intelligence, DeepSeek offers a less scrutinized alternative. Fighters report that when US-based models flag suspicious queries, they pivot to Chinese services—exploiting differences in content moderation and geopolitical oversight.

China’s growing AI ecosystem poses a dilemma for Western intelligence agencies. DeepSeek operates outside Western regulatory frameworks, creating a digital safe haven for groups seeking cutting-edge capabilities without centralized oversight. For European and North American counterterrorism units, this fragmentation complicates surveillance and interception efforts.