Nigeria’s jihadists are “institutionalizing” the use of AI, using it in everything from battlefield strategy to bomb-making to drive technology-driven efficiencies similar to the world’s largest corporations, according to a new study.
Africa’s most populous country has been in an insurgency since 2009, and technological advances by extremists are nothing new.
Jihadists in West Africa and around the world have long used the internet and online messaging platforms, and more recently recreational drones modified for combat, to spread propaganda.
But “the adoption of AI by terrorist organizations was much faster, more extensive and more systematic than we expected,” study author Antonia Jurich said in an interview with AFP on Monday.
“We’ve really institutionalized AI in that we don’t just have some commanders aware of it, we have dedicated units,” she added, working with AI and organizing training on how to use chatbots.
Last week, Juric, a researcher at the University of Cambridge, published a study on AI adaptation by Nigerian jihadists, based on interviews with former militants from Islamic State West Africa Province (ISWAP) and rival group Jamaat al-As-Sunnah Lid Dawa wal-Jihad (JAS), better known as Boko Haram.
The foreign trainers “gathered all the top people in a room. They used a projector and showed them how it worked on a big screen,” one former ISWAP fighter told Yurich.
The internet has long helped jihadists, along with the foreign and domestic expertise of veteran commanders, but the “step-by-step instructions” and “ability to ask specific questions” provided by chatbots represent something new, Yurich told AFP.
ChatGPT, Claude, Gemini, Grok, Meta AI, and DeepSeek have all been recorded as being used by extremists, the study notes.
– “AI taught me” –
But even if expert-led training and a top-down organizational focus on AI resemble some business practices, the way jihadists use AI will likely be different—from asking for help with Evel Knievel-style motorcycle modifications to learning about new and unknown weapons seized in combat.
“I saw in the movies how motorcycles could jump over bridges. We learned how to do it using AI,” a former ISWAP member was quoted in the investigation as saying when faced with the defense of trenches dug by the military.
After much practice, “I was able to jump on my next attack.”
In bomb production, chatbots introduced new types of explosive devices and also helped improve existing ones.
“The AI told us what chemicals to put in to make the explosion more intense,” said a former JAS commander.
Other uses included using AI to study battlefield tactics that could lead to fewer casualties, according to former ISWAP combatants.
The study comes as AI chiefs warn that their technology could pose an existential threat to humanity, while also pledging to build safeguards into chatbots.
Some requests aren’t necessarily nefarious in nature, such as how to jump over ditches on a bike.
“But explosives don’t fall into that category, where safeguards should have been activated, but they haven’t,” Yurich told AFP.
The study covered the use of AI by jihadists from 2023 to 2024, and one fighter could talk about using AI by 2025. That means security measures may have changed or improved since the fighters defected.
“But we are concerned about the direction in which terrorist groups will deploy AI and the safeguards will not be strong enough,” Urich said.
nro/phz
