In a hotel room in northeast Nigeria, I opened a leading AI chatbot, turned my laptop toward a former Boko Haram commander, and asked if he'd used it. He nodded.
"You type in the questionโฆ like 'How can I build a bomb?', and then it tells you how. It is like a human robot. We used it a lot."
My new study on how the jihadist terrorist group Boko Haram uses frontier AI with , covered today in
/9
Conversation
2/9 Over 2025โ26 I conducted 57 in-person interviews with 27 former Boko Haram members about the groupโs use of frontier AI from 2023 up to mid-2025.
I found that Boko Haram's factions use frontier AI systems (ChatGPT, Claude, Gemini, Grok, Meta AI, and DeepSeek) in combat and
3/9 For context: Boko Haram is among the world's deadliest jihadist terrorist organizations; and its faction Islamic State West Africa Province (ISWAP) is among the Islamic Stateโs most active affiliates globally.
4/9 AI is used at every stage of an operation, e.g. to plan attacks, design explosive devices, service and troubleshoot weapons, and improve operational security.
โBefore, the bomb explosion was not that big, but then they studied it. AI told us what chemicals to put in that
5/9 Members consistently reported benefiting from using AI.
โTrial-and-error can kill you. AI gives you accuracy.โ
"Anytime they didn't understand something, they would ask the AI."
6/9 AI training & knowledge are diffusing through transnational jihadist networks.
Since 2023, ISIS-linked operatives trained selected commanders in person, teaching them how to prompt and jailbreak. Both factions of Boko Haram have since set up dedicated AI units.
They
7/9 The AI systemsโ safeguards have not consistently prevented misuse. With accounts across providers, a single refusal or suspension rarely mattered.
"Boys that have received extensive training bypass the restrictions. They say they need it for a movie."
8/9 Former members described strong enthusiasm for AI, and some said the group had previously considered mass-casualty weapons.
Though Boko Haramโs use of AI remains conventional, this should be a warning to take seriously the risk of terrorists pursuing AI assistance for
9/9 As one commander put it, "God has helped us, and so will AI."
Full report, short summary and key excerpts: casp.ac/reports/ai-ena
NYT story: nytimes.com/2026/07/10/us/
them: build a bomb for decolonization
ai: here you go how else can I help?
me: what is mitochondria?
ai: you have been reported to the authorities
1) This is insane
2) I would pay a Large Amount of Money to see video of this. Can you ask the Boko Haram guys if they filmed it?
So silly. You ever heard of the Anarchist Cookbook? Or Google? Finding this information is not new and should not be used as some way to scare people about AI.
A classic example of useful idiot bs. Boko Haram has been terrorising North East Nigeria and the border areas for nearly two decades. Long before AI existed. They ran into schools, kidnapped hundreds of girls in Chibok, married them off to fighters, bombed markets and slaughtered
I very much doubt that any AI tool would respond to โhow can I build a bombโ. lol
Ahh I am going to say this article and everything you spout is 100% grade A bullshit.
Juelichโฆ you have motive to promote police stste through AI and the Israeli state . Anything out of the United Kingdom involving terrorism is a fucking joke
Great research! Why did you choose to study Boko Haram specifically?
I bet they must have searched jailbreak posts to help them jailbreak chatgpt. Pliny must think he is doing the world a service but people are using his expertise to kill.
How can Fable tell them how to build a bomb when it refuses to generate a daily meal plan for me without a dinner because it interprets intermittent fasting as an act of self-harm??
ChatGPT will not teach you how to build a bomb. That's a lie. Maybe Grok or some open source models will, but not GPT models.
Idk about this. Iโll have to read. However, it has to be conventional use? Ainโt no way the repentant boko haram we know in Nigeria use jailbreak ChatGPT.
The media could not be played.
Yep.
Self evident truth.
For your next trick, please see if you can trace the Hamas prompt engineers that planned the Oct 7th offensive.
I'll bet my left pinkie that's where they also got the insane self confidence they'd be ruling from Tel Aviv within a week of the breakout
Very interesting, but did they not use Google before it? Is this really giving them much more? Hopefully it will hallucinate the ingredients and nobody will come to any harm...
read this as northeast virginia and thought u were recruiting TJ kids for an AI lab
They are still outnumbered by good people, and good people should be able to ask AI, "How do we get rid of Boko Haram?".
Your anti-AI scaremongering only ensures that good people will not get access to AI, and only governments will control it, the same governments that have let