# Boko Haram's AI-Driven Terrorist Activity Revealed | Peter Slattery, PhD posted on the topic **By:** Peter Slattery, PhD **Published:** 2026-07-12T18:09:52.139Z **Source:** [LinkedIn](https://www.linkedin.com/posts/peterslattery1_how-the-terrorist-group-boko-haram-uses-frontier-ugcPost-7482134144521629696-2pGo) --- Boko Haram's AI-Driven Terrorist Activity Revealed This title was summarized by AI from the post below. Peter Slattery, PhD 1w Report this post "How are terrorists using AI? Semi-structured interviews with 27 former Boko Haram members conducted in northeast Nigeria in 2025 and 2026 reveal unprecedented detail about AI-assisted terrorist activity primarily through 2024. This report finds that both factions of Boko Haram use frontier AI, including ChatGPT, Claude, Gemini, Grok, Meta AI, and DeepSeek, to assist in combat and day-to-day operations. This AI use is institutionalized through specialized units and internal training. It has aided in attack planning, weapons troubleshooting, and the design of explosive devices, as users have successfully circumvented some safeguards. This know-how was transferred through transnational jihadist networks, with Islamic State operatives delivering in-person training. Respondents expressed strong enthusiasm for AI and, in some cases, openness to mass-casualty weapons, though documented use remains conventional. Terrorist adoption of AI has thus advanced further and more systematically than prior analysis has recognized, making it a present and growing reality that warrants attention from policymakers, security communities, and AI developers." Antonia Juelich at Cambridge Programme on AI Science & Policy (CASP) 113 10 Comments Like Comment Peter Slattery, PhD 1w Report this comment Read/download: https://casp.ac/reports/ai-enabled-terrorism Like Reply 3 Reactions 4 Reactions Ayni Ali S. Q. 1w Report this comment That's genuinely concerning. I was wondering recently whether locally hosted models, such as Ollama, are monitored for malicious use. If reports suggest that even widely used LLMs like ChatGPT and others can still be exploited for harmful planning despite their safety measures, it highlights how challenging AI governance and misuse prevention remain. Strengthening safeguards while preserving legitimate access is becoming increasingly important. Like Reply 1 Reaction German Rodriguez 1w Report this comment AI isn’t just a tool for efficiency; it’s reshaping the entire playbook for these groups. The real danger is how quickly they adapt and evolve tactics—using AI not only for combat but for recruitment, propaganda, and misinformation campaigns. That’s the hidden dimension we need to focus on. The tech isn't the threat; it’s the ingenuity behind it. Like Reply 1 Reaction Kenneth Tannenbaum 1w Report this comment This is exactly the gap observational governance can't close — by the time an interview surfaces this kind of institutionalized misuse, the harm is already done. Execution-time enforcement, not just RLHF-style guardrails, needs to be part of the conversation policymakers are having. Like Reply 1 Reaction Pip Foweraker 🔸 1w Report this comment A heuristic of 'General-purpose tools will tend to find use in the hands of highly agentic people, regardless of their moral positioning' seems reasonable here Like Reply 1 Reaction Stephen Johnson MCSFS 1w Report this comment Great research and (should) be getting regulators and labs thinking very seriously Like Reply 1 Reaction 2 Reactions Broderick James McDonald 1w Report this comment 🎯 Like Reply 1 Reaction Eddy in Holland 1w Report this comment 🌍 After walking an international security research biome spanning AI safety, terrorism studies and field research, the Owl kept returning to the same clearing.🦉The most striking observation wasn't that frontier models were used, it was that:.. AI had already become institutionalized through specialized teams, training, role separation and knowledge transfer. That suggests the organizational adaptation may be as important as the technology itself. 🦉 [p5p1~p7p3] Further along the trail, the Fox paused at another distinction.🦊 Much of the public discussion focuses on individual jailbreaks, while this report points toward coordinated workflows where multiple providers, internal expertise and organizational processes reduce reliance on any single safeguard. 🦊 [p7p4~p9p2] Before leaving this international research forest, the Squirrel tucked away one acorn 🐿️.. perhaps the long-term governance question isn't only how capable an AI model becomes, but how capable an organization becomes at operationalizing AI over time. Thank you for opening another important international research forest. 🌲 Like Reply 1 Reaction See more comments To view or add a comment, sign in More Relevant Posts Jean-Marc Rickli 1w Edited Report this post Not surprising. It was expected but at least now we have some more concrete evidence of #terrorists ’uses of #AI. Prior cases included the use of generative AI for #communication purposes through #deepfakes or #memes, #conversational #AI for recruitment purposes or #ChatGPT for calculating explosive loads. We really have to anticipate and prepare for future uses of #autonomous #swarms for #terrorism purposes not just on the #battlefield! Peter Slattery, PhD MIT AI Risk Initiative | MIT FutureTech 1w "How are terrorists using AI? Semi-structured interviews with 27 former Boko Haram members conducted in northeast Nigeria in 2025 and 2026 reveal unprecedented detail about AI-assisted terrorist activity primarily through 2024. This report finds that both factions of Boko Haram use frontier AI, including ChatGPT, Claude, Gemini, Grok, Meta AI, and DeepSeek, to assist in combat and day-to-day operations. This AI use is institutionalized through specialized units and internal training. It has aided in attack planning, weapons troubleshooting, and the design of explosive devices, as users have successfully circumvented some safeguards. This know-how was transferred through transnational jihadist networks, with Islamic State operatives delivering in-person training. Respondents expressed strong enthusiasm for AI and, in some cases, openness to mass-casualty weapons, though documented use remains conventional. Terrorist adoption of AI has thus advanced further and more systematically than prior analysis has recognized, making it a present and growing reality that warrants attention from policymakers, security communities, and AI developers." Antonia Juelich at Cambridge Programme on AI Science & Policy (CASP) 22 Like Comment To view or add a comment, sign in Yoonsuk Choi 5d Report this post “…How are terrorists using AI? Semi-structured interviews with 27 former Boko Haram members conducted in northeast Nigeria in 2025 and 2026 reveal unprecedented detail about AI-assisted terrorist activity primarily through 2024. This report finds that both factions of Boko Haram use frontier AI, including ChatGPT, Claude, Gemini, Grok, Meta AI, and DeepSeek, to assist in combat and day-to-day operations. This AI use is institutionalized through specialized units and internal training. It has aided in attack planning, weapons troubleshooting, and the design of explosive devices, as users have successfully circumvented some safeguards. This know-how was transferred through transnational jihadist networks, with Islamic State operatives delivering in-person training. Respondents expressed strong enthusiasm for AI and, in some cases, openness to mass-casualty weapons, though documented use remains conventional. Terrorist adoption of AI has thus advanced further and more systematically than prior analysis has recognized, making it a present and growing reality that warrants attention from policymakers, security communities, and AI developers…” Like Comment To view or add a comment, sign in Francesco Virzi' 1w Report this post “God has helped us, and so will AI”: How the Terrorist Group Boko Haram Uses Frontier AI Cambridge has published something that should worry anyone working on AI safety: documented, on-the-ground evidence that a terrorist group is already using frontier AI to fight. Antonia Juelich, at the Cambridge Programme on AI Science & Policy, interviewed 27 former Boko Haram members across 57 sittings in northeast Nigeria. What they described goes well beyond the propaganda use that earlier assessments assumed was the ceiling. Both factions (ISWAP and JAS) have leaned on leading AI systems at every stage of military activity: planning assaults, designing explosives, servicing weapons, tightening operational security. A few things struck me. They built dedicated teams for this. Bomb-makers, gun specialists, and engineers were assigned to query the models and pass strategies down the chain. A resource-starved group spending its best technical people this way says a lot about what it thinks AI is worth. They didn't figure it out alone. Islamic State operatives ran in-person sessions, handed over laptops with VPNs, set up and paid for accounts, and coached the group daily on how to prompt and how to get around platform limits. One former commander's line sticks: "The white guys came and taught us." The guardrails bent. Operators called the restrictions manageable, not prohibitive. They jailbroke with cover stories ("we need it for a movie"), and kept accounts across several providers so one refusal or account suspension changed nothing. Juelich's sharpest point is that none of this depends on Boko Haram being special: it isn't, not in money, not in sophistication nor in motivation. The tools are public, the barriers are low, and one well-evidenced case is enough to call this a problem for right now rather than later. She ends by asking whether AI developers, policymakers, and intelligence agencies have built defenses that match the scale of what she found. The full report can be found here: https://lnkd.in/ekN_P6AU Like Comment To view or add a comment, sign in Syntony 137 followers 1w Report this post "This report finds that both factions of Boko Haram use frontier AI, including ChatGPT, Claude, Gemini, Grok, Meta AI, and DeepSeek, to assist in combat and day-to-day operations. This AI use is institutionalized through specialized units and internal training. It has aided in attack planning, weapons troubleshooting, and the design of explosive devices, as users have successfully circumvented some safeguards. This know-how was transferred through transnational jihadist networks, with Islamic State operatives delivering in-person training. Respondents expressed strong enthusiasm for AI and, in some cases, openness to mass-casualty weapons, though documented use remains conventional. Terrorist adoption of AI has thus advanced further and more systematically than prior analysis has recognized, making it a present and growing reality that warrants attention from policymakers, security communities, and AI developers." Incredible analysis from the Cambridge Programme on AI Science & Policy (CASP). https://lnkd.in/eigrpwvh Like Comment To view or add a comment, sign in Vikas Lohchab 1w Report this post Cons of Artificial Intelligence and we need to counter it with all the might: Field research conducted by Dr. Antonia Juelich of University of Cambridge reveals that Boko Haram has systematically integrated frontier artificial intelligence into its core military and day-to-day operations. Moving far beyond online propaganda, they utilise leading large-language models - including #ChatGPT, #Claude, #Gemini, #Grok, #MetaAI, and #DeepSeek - to secure tactical battlefield advantages. Terrorist operatives consult these chatbots across all phases of military activity, using them to plan ambushes, troubleshoot jammed weaponry, design sophisticated explosive devices, and optimise force allocation. In one instance, fighters used #AI -derived specifications to modify motorcycles and master trench-jumping techniques to breach fortified military bases. Furthermore, AI assists in post-mission learning and training media teams to evade digital surveillance. Rather than relying on casual, isolated usage, both factions have institutionalised AI by establishing dedicated technical units supervised by senior leadership. Transnational jihadist networks facilitated this rapid adoption; Islamic State operatives provided laptops, managed paid subscriptions, and delivered in-person training on prompt engineering and jailbreaking protocols. Consequently, built-in corporate safeguards have proved manageable rather than prohibitive, with users easily circumventing safety restrictions via basic coaxing methods. This systematic adoption underscores a critical vulnerability: motivated, resource-constrained adversaries can easily exploit publicly available AI to uplift conventional capabilities. https://lnkd.in/grGK_t-M 5 Like Comment To view or add a comment, sign in Rina Chandran 1w Edited Report this post I’ve heard the term “flourishing” so often in relation to AI recently, and would like to believe that everyone is working on AI for our good. But it’s naive to assume AI is only being used for benign and useful purposes. Two new reports show that extremist groups are using AI extensively, and that AI is likely to reinforce lone-actor violence, digital propaganda, and information warfare. A study by Antonia Juelich found that Boko Haram is using frontier AI, including ChatGPT, Claude, Gemini, Grok, Meta AI, and DeepSeek, to assist in combat and day-to-day operations. Until recently, extremists primarily used AI for propaganda, translation, recruitment and fund raising. But they now turn to chatbots to design explosives, fix or upgrade other weapons, and brainstorm ideas on how to attack their enemies. Boko Haram has set up dedicated AI teams, and AI training is being done through transnational networks. While chatbots have some safeguards to prevent users from soliciting information that could cause harm to others or themselves, researchers have repeatedly found that people can circumvent the safety protocols. “The terrorists are not waiting for us to make AI safe,” Dr. Juelich told the New York Times, adding that their use of AI had been “significantly underestimated in both scope and character.” A separate report by Daniel Byman and V.S. Subrahmanian notes that AI lowers barriers, optimizes workflows, enables personalization of propaganda and recruitment, and allows smaller groups and individuals to perform tasks that have previously required more manpower, technical expertise, time, and money. AI will also strengthen state counterterrorism capabilities through improved surveillance, pattern recognition, and intelligence analysis, they say. But that also means that civilians have to contend with greater surveillance. Read the reports here: https://lnkd.in/gRarE5zp https://lnkd.in/g9KKQsfq 4 Like Comment To view or add a comment, sign in Nop.ng 726 followers 1w Report this post Report: Boko Haram Used AI Chatbots To Build Bombs, Plan Attacks A report by The New York Times has revealed how Boko Haram terrorists turned to artificial intelligence (AI) for weapons making, including bombs, as well as improving their operations generally. Former insurgents, the report noted, used multiple AI platforms, including ChatGPT, Claude, Gemini, Grok and DeepSeek. According to the report, former Boko Haram commanders told terrorism and technology researcher Antonia Juelich of the University of Cambridge that after an assault on a military base was frustrated by a defensive trench, the group consulted AI tools for ideas.... Like Comment To view or add a comment, sign in TopStories Nigeria 22 followers 1w Report this post Report: Boko Haram Used AI Chatbots To Build Bombs, Plan Attacks A report by The New York Times has revealed how Boko Haram terrorists turned to artificial intelligence (AI) for weapons making, including bombs, as well as improving their operations generally. Former insurgents, the report noted, used multiple AI platforms, including ChatGPT, Claude, Gemini, Grok and DeepSeek. According to the report, former Boko Haram commanders told terrorism and […] Full Details: https://lnkd.in/gvnHHA9Z Like Comment To view or add a comment, sign in TheOutpost.ai 593 followers 1w Report this post A University of Cambridge study has uncovered something the AI industry can no longer ignore: terrorist groups are not just experimenting with AI — they've built structured systems around it. Based on 57 in-person interviews with 27 former Boko Haram members in Nigeria, researcher Antonia Juelich found that groups like Boko Haram and the Islamic State are using tools like ChatGPT, Gemini, Claude, Grok, Meta AI, and DeepSeek for: • Bomb-making instructions and weapons troubleshooting • Tactical attack planning and day-to-day operations • Recruitment and propaganda generation • Internal AI training programs with dedicated units 🔍 What makes this especially alarming: This wasn't isolated misuse. These groups have established dedicated AI teams, structured knowledge-sharing, and even remote AI training assistance — mirroring how legitimate organizations adopt new technology. One former member described the direct impact: "Before, the bomb explosion was not that big, but then they studied it. AI told us what chemicals to put in that made the explosion heavier." Safety filters? Bypassed easily — recruits simply told chatbots they needed the information "for a movie." 💬 What are tech companies saying? OpenAI and Meta have acknowledged the issue, stating they've strengthened safeguards in newer model versions. OpenAI's Drew Pusateri noted: "We know that bad actors will never stop trying to misuse our tools, and we'll continue strengthening our defenses." But researchers caution that the challenge is deeply complex. Gray areas — like asking about chemistry or vehicle repair — make blanket restrictions nearly impossible without limiting legitimate use. ⚠️ The bottom line from the researcher: "The terrorists are not waiting for us to make AI safe. They are able to use them now and train them to cause harm." This is a critical moment for AI governance, policy, and safety design. The industry needs to move faster. 📖 Read the full story here: https://lnkd.in/esBYQNb9 #AISafety #AIPolicy #ArtificialIntelligence #CyberSecurity #TechEthics #NationalSecurity #ResponsibleAI #AIGovernance Like Comment To view or add a comment, sign in 70,837 followers 1,279 Posts View Profile Connect More from this author Mistakes in audience research (part 2) Peter Slattery, PhD 4y The READI philanthropy research database: September 2021 update Peter Slattery, PhD 4y What Influences the consumption of Animal-Products? A Meta-Review Peter Slattery, PhD 4y Explore related topics How Hackers Use AI in Cyber Attacks ChatGPT Data Security Risks Using AI for Personal Insight Analysis How Developers can Use AI in the Terminal How Governments Use AI to Combat Financial Crime How to Use AI Instead of Traditional Coding Skills AI Training for Cybersecurity Engineers Advanced Analytics Strategies Using AI Explore content categories Career Productivity Finance Soft Skills & Emotional Intelligence Project Management Education Technology Leadership Ecommerce User Experience