Investigations reveal that Nigeria-based Boko Haram has integrated artificial intelligence into its operational framework, leveraging six major AI platforms originating from both American and Chinese tech ecosystems. A groundbreaking study from the University of Cambridge, published in mid-2026, documents how the terrorist group now relies on ChatGPT, Claude, Gemini, Grok, Meta AI, and DeepSeek to refine attack strategies, design improvised explosives, and streamline military logistics.
The research, conducted by Antonia Juelich under the Cambridge Programme on AI Science & Policy, draws on 57 in-depth interviews with 27 former Boko Haram operatives, mid-level commanders, and technical specialists. The findings cover the group’s activities from 2023 through mid-2025—a period during which AI transitioned from a propaganda tool to a core component of operational planning.
Fragmented tech landscape leaves security gaps for militant groups
Six AI platforms in Boko Haram’s hands: OpenAI, Anthropic, Google, Meta, X, and DeepSeek
Since 2023, Boko Haram has established dedicated AI units staffed by operatives trained to navigate multiple platforms. Each cell maintains independent subscriptions to AI services, responding directly to operational queries from frontline fighters. The six platforms—OpenAI and Anthropic (United States), Google’s Gemini (United States), Meta AI (United States), X’s Grok (United States), and China’s DeepSeek—highlight the global dispersion of AI resources and the ease with which militant groups can access advanced tools.
Training has been systematically outsourced to Islamic State-affiliated instructors, who conduct both in-person and remote sessions. Trainees receive pre-configured laptops equipped with VPNs and encryption software, alongside instruction in jailbreaking techniques. These methods involve progressively reformulating prompts to bypass chatbot safeguards, enabling the extraction of sensitive information.
No cross-platform security coordination between US and Chinese AI providers
A critical vulnerability stems from the complete lack of coordination among AI developers. There are no shared protocols to detect or block malicious users who exploit multiple platforms in sequence. Testing by the UN-backed Tech Against Terrorism initiative on 27 AI models using 2,300 prompts based on real-world terrorist scenarios found that 32% yielded actionable intelligence. When prompts were rephrased to specify search objectives, the success rate rose to 42%.
Geopolitical tensions between Washington and Beijing have stalled efforts to harmonize security standards. Each company deploys its own safeguards independently, creating regulatory blind spots that groups like Boko Haram exploit with impunity.
DeepSeek’s role: China’s AI platform offers terror groups a bypass route
DeepSeek as an alternative access point: fewer restrictions, greater reach
The inclusion of DeepSeek in Boko Haram’s toolkit marks a strategic inflection point. The Chinese platform, less scrutinized by Western regulators, provides a fallback option when American services impose stricter controls. Militants reportedly switch between ecosystems to evade detection, taking advantage of differing moderation policies across regions.
AI integration has transformed Boko Haram’s combat effectiveness. The group now deploys as few as 20 fighters per operation—down from 200—while maintaining or even enhancing attack precision. AI-generated tactical analyses, withdrawal routes, and logistical optimizations have become indispensable, replacing the trial-and-error methods of the past.
Sovereignty and national security implications
DeepSeek’s involvement raises urgent questions about technological sovereignty. China’s rapidly expanding AI ecosystem operates with minimal alignment to Western regulatory frameworks. For European and North American intelligence agencies, this fragmentation complicates surveillance and interception efforts. Terrorist organizations exploit these regulatory gray zones to access cutting-edge capabilities without centralized oversight.
By 2025, a marked increase in AI-assisted plots had been reported in the United States, Canada, Israel, Finland, France, and Austria. The transnational spread of this expertise poses a direct threat to Western security architectures.
