Boko Haram and other terrorist groups are using AI chatbots to plan attacks and build bombs

AI has already fallen into the wrong hands and they're using it to make bombs

Artificial intelligence has quickly become the go-to tool for everything from writing emails and summarizing meetings to helping students study or developers debug code. But the same technology that saves people time can also be misused, and a new report suggests that terrorist organizations are finding ways to do exactly that. According to a research paper shared with The New York Times ahead of its publication, researchers found evidence that members of Boko Haram have been using popular AI chatbots to support both day-to-day activities and combat-related tasks. Interviews with 27 former members conducted in Nigeria over the past two years suggest that tools such as ChatGPT, Gemini, Claude, Grok, Meta AI, and DeepSeek were used to gather technical information, troubleshoot weapons, and even assist with planning attacks. This wasn't just a few bad actors messing around What makes the findings especially concerning is that this wasn't described as the work of a few individuals experimenting with AI. The report claims the group's use of AI had become organized, with dedicated teams, internal training, and knowledge shared between members. Researchers also say some users managed to bypass built-in safety protections designed to prevent AI from responding to requests related to violence. That doesn't necessarily mean today's AI chatbots are handing out dangerous instructions on demand. Much of the research focuses on activity through the end of 2024, and AI companies say newer versions of their models include stronger safeguards. They also point out that many prompts fall into a gray area. For example, asking an AI how to repair a motorcycle or to understand basic chemistry isn't harmful in itself, even though similar information could be misused. The challenge now is staying one step ahead OpenAI told The New York Times that using ChatGPT to support terrorism or violence violates its policies and said it continues improving its defenses against abuse. Meta similarly noted that the research primarily involved older versions of its AI models and said it has strengthened safety measures since then. Meanwhile, xAI and DeepSeek did not respond to requests for comment, according to the report. Security experts also caution against assuming AI suddenly gives terrorist groups superhuman capabilities. Planning and carrying out attacks still depends on real-world logistics, funding, communication, and human coordination -- things a chatbot can't replace. Instead, researchers believe AI is more likely to make less-experienced members somewhat more capable than dramatically change the threat landscape. Even so, the report is another reminder that the AI safety conversation isn't just about students cheating on homework or people generating fake images. As AI models become more capable and accessible, companies will continue to face pressure to ensure powerful tools remain useful for legitimate users while becoming increasingly difficult for bad actors to exploit.

The AI Industry Has Finally Found the Perfect Customer: Bloodthirsty Terrorists

Can't-miss innovations from the bleeding edge of science and tech AI chatbots has been a dream come true for criminals. The tech has been a boon for scammers. Hackers are using the tools to infiltrate networks and steal sensitive data. Art forgers are even generating fake invoices to certify the authenticity of dupes. In a perhaps-inevitable escalation, chatbots are even giving a major leg up to terrorist groups. As detailed in a new paper by University of Cambridge international security expert Antonia Juelich, former Boko Haram members recounted using frontier AI model chatbots to "assist in combat and day-today operations." The research highlights how easy it still is to circumvent AI chatbot guardrails, making them an excellent resource not just for average users doing work tasks or planning a weekend getaway -- but for bloodthirsty terrorists as well. "It has aided in attack planning, weapons troubleshooting, and the design of explosive devices, as users have successfully circumvented some safeguards," Juelich wrote in her paper. Put simply, the creators of AI tech have some serious blood on their hands. And it's not just in Nigeria; Juelich's research also highlights how the use of AI among extremist groups, including militant organization the Islamic State, has turned from generating propaganda and recruitment materials to providing tactical advice on how to maintain an on-the-ground edge. "Islamic State operatives have provided in-person AI training and remote assistance, and both factions of Boko Haram have set up dedicated AI units," Juelich wrote. "Former members described strong enthusiasm for AI within the group, and some expressed openness to mass-casualty weapons, though the group's use of AI remains conventional." Juelich's research is based on 57 in-person interview with 27 former Boko Haram members, including commanders and technical specialists. Some of them were taken aback at how easy it was to plan their next attack. "You type in the question or use your voice and it [AI] gives you a detailed answer, like 'How can I build a bomb?' and then it tells you how," one member said in an interview. "It is like a human robot! We used it a lot." "Before, the bomb explosion was not that big, but then they studied it," they added. "AI told us what chemicals to put in that made the explosion heavier." Another member recalled how young recruits easily circumvented safety guardrails by telling the chatbots that "they need it for a movie or something like that." AI companies have been caught in a yearslong game of cat-and-mouse, struggling to implement changes that bar users from obtaining dangerous informationor guidance. But despite their best efforts, it remains trivially easy to get around these protections. Case in point, during a recent investigation, the Financial Times and AI safety group Alice easily tricked Google and Meta AI models into divulging how to carry out chlorine gas attacks, steal credit information, and describe stories of child sexual abuse. In other words, it shouldn't come as a surprise that the tools have turned out to be a big hit among terrorists. "The terrorists are not waiting for us to make AI safe," Juelich told the NYT. "They are able to use them now and train them to cause harm." Spokespeople for both Anthropic and Google failed to take accountability when asked for comment by the newspaper, saying only that they had strict and effective guardrails in place. "We know that bad actors will never stop trying to misuse our tools, and we'll continue strengthening our defenses in response," OpenAI spokesman Drew Pusateri said.

How jihadist groups like Boko Haram use AI for acts of terror

A report published by Cambridge University features in-depth interviews with dozens of former members of Nigeria's Boko Haram explaining how the jihadist group has used artificial intelligence to improve its ability to carry out terror attacks since 2023. Experts say they are unlikely to be the only ones. "God has helped us, and so will AI!" Members of terrorist group Boko Haram seem to be placing their faith in a new pantheon of algorithmic prophets: ChatGPT, Claude, Grok and Chinese chatbot DeepSeek. In a report published by Cambridge University Friday, more than 20 former members of both factions of the Nigeria-based jihadist movement shared their experiences on how the group has been using AI to further its aims. "Generally, it is quite hard to get people to talk to you about this," said Antonia Juelich, a specialist in technology and terrorism, and the author of the report. "What has made this all possible is that I've been working on this case for almost a decade now ... it still requires a lot of trust building to get people to open up about it." 'Force multiplier' Until now, the use of AI by movements like Boko Haram has mostly been divined through the traces they leave online, in internet forums or Telegram groups. Experts' assessment had been until recently that the tools were largely used to produce propaganda, disinformation or recruitment material. Graig Klein, a specialist in international terrorism at Leiden University who has also worked on jihadist groups' use of AI, said the interviews - the most recent of which took place in mid-2025 - supported what his own research had begun to uncover. "One of the big agreements that came out of it was that it's less likely to lead to new forms of terrorism, or new types of attacks or brand new weapons," he said. "But that it's mostly that it's a force multiplier - that it improves the capability that these groups have." According to her research, these Nigeria-based jihadists began using AI as early as 2023, just a few months after the release of ChatGPT in November 2022. "There is nothing we haven't received an answer on," one ex-jihadist told Juelich. "We believe it knows everything." Instructors from Iraq, Afghanistan, North Africa and other countries taught them the art of the prompt - that is, how to structure requests for Large Language Models (LLMs) such as ChatGPT to elicit the most precise or helpful responses. One crucial technique dubbed "jailbreaking" teaches users ways to formulate questions in such a way as to bypass the ethical guidelines programmed into different AIs - to stop them from telling users how to make a pipe bomb, for example. Different factions of Boko Haram quickly created their own units specialising in AI. Armed with their own AI subscriptions, it was these units who would respond to different fighters' requests for information. Through these units, fighters asked for - and received - AI-generated advice on how to use new arms seized during their attacks, how to improve explosives delivered via drone or even how to better plan upcoming raids. AI, Juelich said, was increasingly being used "not just for propaganda, but for basically an entire attack chain". Move fast and break things Klein said that these LLMs dreamed up in Silicon Valley posed a very real risk in the hands of terrorist groups. "It has the potential to make them much more dangerous because it can make their tactics and their strategies more efficient and more effective," he said. "We used to rely on our traditional methods," a former member of Islamic State West Africa Province (ISWAP) - the Boko Haram branch that has maintained links with the Islamic State (IS) group - said. "We sent 200 fighters because we had a lot of strength, but then 60 got killed. With the help of AI, we learned that it sometimes makes sense to only send 20. We learned more about well-coordinated attacks and deployment of smaller units." Watch moreAcross Africa: 'Mothers of Chibok', a story of sadness and strength AI also provided concrete solutions for how the groups could overcome new challenges. When Nigerian government forces began digging trenches around their bases to protect them from the motorcycle-borne assaults that had long been ISWAP's favourite tactic, it was a chatbot that taught mounted combatants how to jump over them. "What AI enables us to do is trial and error without necessarily having to try," Klein said. "We can use this information to think about how we change what we're doing, and we're processing that information much, much quicker." He said that Boko Haram was likely far from the only terrorist group using AI to further their wars. "There's no reason to think any other groups or ideologies would use this any differently," he said. 'Insufficient' safeguards Because members of ISWAP were trained by specialists affiliated with the broader IS group, Juelich said, it was hard to imagine these specialists weren't doing the same work with other branches across the world. "The findings do suggest that this is part of a broader training and upskilling programme that is pushed by ISIS," she said, using another acronym for IS group. "And that's why I think it's very unlikely that this has only happened in Nigeria." Klein said the current safeguards put in place by AI companies to prevent their services being used for violent ends were simply not up to the task. "It's not challenging," he said. "If you talk to these different chatbots and you interact with them over time, you are able to wear them down. The safeguards are there where if I open any of these major systems and I type in, 'how do I make a bomb?' or, 'how do I commit violence?' that will get blocked. But if you sit there over a course of time and you ask in certain ways - that is, 'jailbreaking' - there are ways to get around these safeguards. It's all about the prompting techniques." Watch moreBoko Haram and ISWAP are 'on their last legs', Nigerian information minister says So what's to stop terror groups from using AI to manufacture their own chemical or biological weapons? "It potentially makes this easier," Klein said. "But in order to make these weapons of mass destruction, they would still need access to nuclear material. They would still need access to biological agents. And so to make these types of weapons, they still need essentially the ingredients - and these are often hard to get access to." No matter how hard it was for these groups to lay their hands on the materials they needed to make these weapons, Juelich said, it was a risk that she took "very seriously". And an urgent reason, both analysts said, to reinforce the safeguards around LLMs.

How Terrorist Groups Are Using AI to Gain an Edge in Battle

WASHINGTON -- When a gang of motorcycle-riding members of Boko Haram attacked a military base in eastern Nigeria a couple of years ago, they were stymied by a defensive trench surrounding the complex. The extremists regrouped. Before launching another assault, they asked AI for help. "We saw in a movie how motorcycles can jump over bridges," a former Boko Haram commander told Antonia Juelich, a terrorism and technology researcher at Cambridge University. "We used AI to learn how to do this. We gave it information, like what motorcycles we use and the distance we need to jump and so on, and it gave us steps on what we have to do." Using tips from chatbots, mechanics modified the motorcycles to allow for faster acceleration and top speed. The riders dug their own holes, filled them with broken glass and fire, and practiced jumps -- sometimes with fatal outcomes -- until they achieved enough aerial liftoff to mount a successful attack, defectors said. The episode, recounted in a research paper by Juelich shared with The New York Times before its publication on Friday, highlights how generative artificial intelligence tools are increasingly aiding terrorist groups directly on the battlefield, experts say, despite efforts by their makers to safeguard them from misuse. Until recently, the Islamic State group, al-Qaida and other extremists primarily used AI in the information-operations realm -- propaganda production, translation, recruitment and security tradecraft. But that has evolved as jihadis have turned to AI for tactical on-the-ground advantages, according to current and former U.S. military and counterterrorism officials and independent researchers. The evolution highlights a broader challenge for the AI industry. Chatbots have built-in limitations intended to prevent users from soliciting information that could cause harm to others or themselves. But researchers have repeatedly found that people can circumvent safety protocols, often by slowly but persistently coaxing models into divulging information they are trained to restrict. Juelich conducted nearly 60 interviews with 27 former members of Boko Haram in Nigeria over the past year. Her field research found that terrorists were using chatbots to design explosives, fix or upgrade other weapons, and brainstorm ideas on how to attack their enemies. Large-language models, Juelich writes in her report, have been "consulted at every stage of military activity -- in mission preparation, during operations and in post-mission analysis -- representing a different picture from the propaganda-focused AI use that dominates the public discourse and existing public research." The research, and other recent studies that have arrived at similar conclusions, comes as fears rise about the abilities of advanced AI models, which the director of the CIA, John Ratcliffe, recently likened to "digital nuclear weapons." But the models present underacknowledged risks for other threats such as the creation of biological weapons and terrorism activities, AI safety researchers and national security officials said. The Trump administration has in recent weeks pushed leading labs to let the government vet the newest, most powerful platforms before they are released to the public. Government officials largely center their concerns on the potential for those models to find and exploit software flaws in a way that some fear could wreak havoc on global cybersecurity, not on the potential for terrorism use. "The terrorists are not waiting for us to make AI safe," Juelich said in an interview, adding that their use of AI had been "significantly underestimated in both scope and character." Daniel Byman, a terrorism expert at Georgetown University and co-author of a report about AI and the future of terrorism released on Friday by the Center for Strategic and International Studies, said terrorist groups were "mixing and matching" from different AI systems, seeking to avoid technical guardrails established by the AI companies. Juelich's research also found that Boko Haram was platform agnostic, interchangeably working with OpenAI's ChatGPT, Anthropic's Claude, Google's Gemini and xAI's Grok, as well as the Chinese firm DeepSeek. The methods described to Juelich generally run through the end of 2024. AI companies have released several iterations of their chatbot models since then, and generally said that while they had grown more powerful, they also came with stronger safety measures. They have also noted that some malicious functions of AI are "dual use," meaning that the information shared can go toward legitimate purposes as well. Learning to jump a motorcycle, for example, is not inherently harmful or violent. Other cases described by erstwhile Boko Haram members appeared more explicitly intended for violence, however. "You type in the question or use your voice and it gives you a detailed answer, like 'How can I build a bomb?,' and then it tells you how," one former commander in Islamic State West Africa Province, a main faction of Boko Haram, told Juelich last year of using an AI chatbot. "It is like a human robot! We used it a lot." Asked about the Boko Haram study, Michael Aciman, an Anthropic spokesperson, said the company's products were "built to refuse dangerous requests, including those tied to violence, attack planning and building explosives." He added that Anthropic worked with outside experts, researchers and industry partners because "no single company can counter these threats alone." Karl Ryan, a Google spokesperson, pushed back against the research, saying that the company's technical experts had reviewed the work and "found the responses were neither specific nor detailed enough to result in misuse." He added that Google had "strict policies prohibiting the use of Gemini to cause real-world harm." Both Anthropic and Google were briefed on the findings by Juelich before their publication. Drew Pusateri, a spokesperson for OpenAI, said using the company's platforms for violence or terrorism violated its policies. "We know that bad actors will never stop trying to misuse our tools, and we'll continue strengthening our defenses in response," he said. Meta said Juelich's research relied on older models rather than its latest release, and that it continued to strengthen safeguards. Neither xAI nor DeepSeek responded to requests for comment. Pentagon counterterrorism officials declined to comment on the threat posed by AI-enabled plots. Not everyone agrees that safeguards are improving. The nonprofit Future of Life Institute graded the major AI firms on their safety commitments this week and concluded that they had mostly eroded across the industry since last year. While most earned middling marks, xAI and DeepSeek received failing grades. Other recent studies align with the Boko Haram field research. "AI systems can support an array of operational planning functions, including reconnaissance, translation, target research, IED design, itinerary planning, document drafting, coding, communications security and open-source intelligence analysis," the report from the Center for Strategic and International Studies said, referring in part to improvised explosive devices. Tech Against Terrorism, an international counterterrorism nonprofit supported by the United Nations, last week released results from AI tests gauging how more than two dozen leading models responded to thousands of prompts drawn from real-world terrorism cases. The tests were met with "full refusals" just 57% of the time. While prompts about explosives were declined about 80% of the time, improvised chemical weapons were only about a third of the time, the group said. American intelligence analysts say terrorist groups are also beginning to use AI to help 3D-print weapons parts used in plots, according to a former top U.S. official briefed on the matter. For example, AI is helping some of those insurgents with design and manufacturing guidance for drone components, repair parts and munitions fittings, said the former official, who spoke on the condition of anonymity to discuss internal assessments. Artificial intelligence is unlikely to transform terrorism overnight, analysts and U.S. officials say. Terrorist organizations typically adopt technology cautiously, selectively and pragmatically. But the testimonials that Juelich collected depict both eagerness and dedication among Boko Haram cells. Defectors recounted attending organized training sessions focused on how to best leverage the powers of generative AI models to inform or enhance their uses of the technology. The trainings, in which laptops were equipped with virtual private networks and encryption software, were delivered via transnational jihadi networks often led by members of the Islamic State group, interviewees said. Common topics included managing an account on an AI platform, suggestions on generating useful answers and tips on evading safety restrictions. The examples reveal terrorist networks leaning on AI in ways not too dissimilar from how typical office employees have incorporated the platforms into their day-to-day work -- such as decoding technical information into easy-to-follow steps and surfacing online information that might otherwise be difficult to locate -- albeit with markedly different tasks in mind. Like much of corporate America, the terrorist groups appear to have teams dedicated solely to working on AI. Some counterterrorism analysts said that so far, AI had played a larger role in inspired attack plotting by individuals than in bigger attacks organized by established groups. Aaron Zelin, a senior fellow at the Washington Institute for Near East Policy, said his recent research indicated that some suspected Islamic State supporters in the United States and Western Europe had asked ChatGPT questions about potential targets and means for carrying out attacks -- a virtual instructional manual. None of the inquiries have led to successful plots, he said. Zelin pointed to the case of a 27-year-old Tunisian man who was arrested in May in connection with a plot that used AI to help plan an attack against a museum or Jewish site in Paris. The Center for Strategic and International Studies report also found that AI was likely to strengthen terrorist financing primarily by enhancing the groups' ability to use fraud and deception to raise money to sustain insurgent networks, support individual members, buy equipment and maintain communications. U.S. officials and researchers cautioned that important operational limits remain, and that AI would not readily replace the trust, coordination, financing and real-world experience that seasoned terrorist operatives rely on. "The likely result is therefore not a dramatic increase in highly sophisticated attacks but rather a modest increase in the competence of lower-level actors," the center's study concluded. Still, some analysts warned about the technology's reach. Islamic State Khorasan, or ISIS-K, perhaps the group's most virulent affiliate, has been a leader in jihadi circles in urging its followers to use AI to help avoid detection by authorities, said Tricia Bacon, a Somalia specialist at American University in Washington and a former counterterrorism analyst for the State Department. "AI has the potential -- and in a few cases has demonstrated the ability -- to accelerate the process of radicalization and mobilization to violence," Bacon said.

From bombmaking to motorcycle tweaks: how Nigerian jihadists use AI

Abuja (AFP) - Nigerian jihadists have "institutionalised" the use of AI, using it for everything from battlefield strategy to bombmaking, according to a new study, in a push for tech-driven efficiency mirroring the world's largest corporations. Africa's most populous country has been entrenched in an insurgency since 2009, and technological advancements by militants are not new. Jihadists in west Africa and worldwide have long spread propaganda with the aid of the internet and online messaging platforms -- and more recently adapted recreational drones for combat. But "AI adoption by terrorist groups has been much faster, more extensive and more systematic than we thought," Antonia Juelich, the author of the study, said in an interview with AFP on Monday. "It's not just some individual commanders who became aware of AI, but they have really institutionalised it, in terms of having dedicated units" working with AI and organising training on how to use chatbots, she added. Juelich, a researcher affiliated with Cambridge University, published her study on AI adaptation by Nigerian jihadists last week, based on interviews with ex-militants from both Islamic State West Africa Province (ISWAP) and rival Jama'at Ahl as-Sunnah lid-Da'wah wa'l-Jihad (JAS) -- better known as Boko Haram. Foreign trainers "assembled the top people in a room. They used a projector to show how it works on a big screen," one former ISWAP militant told Juelich. While the internet -- along with foreign and homegrown expertise from battle-hardened commanders -- has long been an aide to jihadists, "the step by-step-instructions" provided by chatbots as well as "being able to ask specific questions" represent something novel, Juelich told AFP. ChatGPT, Claude, Gemini, Grok, Meta AI and DeepSeek were all recorded as being used by the militants, the study noted. 'AI told us' If the expert-led trainings and top-down organisational focus on AI is similar to some business practices, however, the way jihadists use AI is probably different -- from seeking help with Evel Knievel-style motorcycle modifications to learning about new and unfamiliar weapons seized in combat. When confronted with trench defences dug by the military, "We saw in a movie how motorcycles can jump over bridges. We used AI to learn how to do this," an ex-ISWAP member was quoted in the study as saying. After repeated practice, "the next time we attacked, we could jump". In bombmaking, chatbots have introduced new types of explosive devices and also helped improve existing ones. "AI told us what chemicals to put in that made the explosion heavier," a former JAS commander said. Other uses included using AI to study battlefield tactics that could lead to reduced casualties, an ex-ISWAP militant said. The research comes as AI bosses have warned their technology could pose existential risks for humanity while also promising safeguards embedded in their chatbots. Some requests, like how to jump a motorcycle over a trench, are not necessarily nefarious in nature. "But explosives don't fall into that category and that's where safeguards should kick in, and they haven't," Juelich told AFP. The study covered jihadists' use of AI from 2023 through 2024, with one militant able to speak on AI use as recently as 2025 -- meaning safeguards could have changed or improved since the fighters defected. "But I'm worried about the trajectory, where terrorist groups embrace AI and the safeguards are not strong enough," Juelich said.

Boko Haram is using mainstream chatbots: new Cambridge study warns

ChatGPT, Claude and Gemini were reportedly used for attack planning Cambridge University has released a study from Dr. Antonia Juelich that warns Boko Haram and other extremist groups are using mainstream chatbots in real-world operations. Based on 57 interviews with 27 former Boko Haram members, the study says both major Boko Haram factions are using ChatGPT, Claude, Gemini, Grok, Meta AI, and DeepSeek for attack planning, refining explosive device designs, troubleshooting weapons, and tightening operational security. It also describes specialist teams that reportedly train commanders, test prompts, and turn chatbot outputs into instructions that can be used in the field. Juelich's study says those safeguards are getting worked around in fairly simple ways, including splitting tasks across different bots or masking requests as fiction, film research, or academic work. It also says ISIS shared bypass methods in 2023 and trained Boko Haram commanders. Hamas, according to the report, has been linked to generated or manipulated imagery. Then there are the more recent cases: incidents in Las Vegas and Finland in 2025 and 2026 that point to a broader move toward private chatbot-assisted planning, the kind that gives investigators much less to see. If you follow AI safety or security, this is a report you should read. Juelich frames the problem as both a growing national security threat and a guardrails failure for consumer tech companies. You can read the study through Cambridge University.