Boko Haram and ISWAP Use AI in 2026 — Lake Chad Changes Doctrine

In Maiduguri, in Borno State, it is not a spectacular drone or a captured armored vehicle that reveals the next tactical breakthrough. It is a training room, a projector, and fighters learning to query ChatGPT, Gemini, Claude, Grok, Meta AI, or DeepSeek. According to an AFP survey published by France 24 on July 13, former members of the Islamic State in West Africa (ISWAP) and Jama'at Ahl as-Sunnah lid-Da'wah wa'l-Jihad (JAS), better known as Boko Haram, describe a faster and more structured adoption of artificial intelligence than expected.

The details matter. We are not just talking about a young propagandist playing with a chatbot. The study cited by AFP, conducted by Antonia Juelich, a researcher affiliated with Cambridge, mentions dedicated units, organized training, and commanders capable of integrating AI into very concrete field problems. For an insurgency that began in 2009 in northeastern Nigeria, this seems less like a technological curiosity and more like a new layer in the tactical toolbox.

An Innovation Grounded in the Field

The reported examples are disturbing because they are mundane. A former ISWAP fighter explains that his group used AI to understand how to get motorcycles across military trenches after seeing jumps over bridges in a movie. In Borno, a motorcycle is not a folkloric detail. It is the most suitable means of reconnaissance, raiding, retreat, and liaison for the bush, sandy tracks, and the shores of Lake Chad.

On the ground, a trench does not stop an ideology. It stops a front wheel. If fighters learn to modify motorcycles, repeat the movement, and choose a shorter angle of attack, the defensive device loses some of its effectiveness. This is not science fiction. It is light infantry solving an obstacle problem with a digital tool.

The other aspect is more serious. AFP reports that former JAS members described the use of chatbots to improve explosive devices, with instructions on chemicals that could enhance the explosion's effect. The precise details should not be reproduced, but the operational implication is clear: the safeguards of the platforms have not always prevented military-purpose research. Again, AI does not replace the explosives specialist. It allows them to query, compare, and test hypotheses faster.

ISWAP and JAS Do Not Use AI in the Same Way

It is important to distinguish between the two organizations. ISWAP, which emerged from Boko Haram's pro-Islamic State split, has often demonstrated superior administrative discipline, better management of local taxes, and a greater ability to combine propaganda, intelligence, and selective violence. If AI is institutionalized in this framework, it can be used for training, translation, military video analysis, recruitment messages, and raid preparation.

The JAS, more fragmented and brutal in its relationship with civilian populations, may find something else: ad-hoc assistance for manufacturing, repairing, and improvising. In both cases, the most likely effect is not an immediate revolution, but a reduction in the cost of learning. A group that loses an expert can train another one faster. An isolated detachment can pose a technical question without waiting for a trainer from another area.

It is new, but it is not magic.

The testimonies remain those of defectors. They do not say how many fighters are trained, nor if the responses obtained are always usable. Some may be exaggerated, as former militants quickly understand what interests researchers and journalists. But the number of tools mentioned and the concrete nature of the uses make the alert credible.

Lake Chad Becomes a Discreet Laboratory

The operational theater of Lake Chad is ideal for this type of experimentation. The borders between Nigeria, Niger, Chad, and Cameroon are easier to trace on a map than to control on the ground. The islands, seasonal swamps, and smuggling routes allow cells to test a technique in one area and then move it elsewhere. For years, the armies in the region have been chasing this mobility.

AI adds a less visible dimension than the commercial drones observed in other African conflicts. A drone leaves debris, a video, a signature. AI training leaves few traces. It can take place in a house, with a phone and an intermittent connection. This is precisely what makes the threat difficult to measure.

On July 14, UN News headlined that terrorism in West Africa and the Sahel "is changing its face." This phrase is accurate under one condition: not to forget that the face rarely changes all at once. In the Sahel as in the Lake Chad basin, armed groups mix the old and the new. The motorcycle, the ransom, local preaching, Telegram video, quadcopter drone, and now the chatbot belong to the same continuum.

What Armies Need to Watch

The first response will not be a miracle software. Nigerian, Cameroonian, Nigerien, and Chadian forces must first monitor material indicators: repeated modifications of motorcycles, new profiles of explosive devices, crossing tactics, instructions typed on captured phones, training materials found after an operation. Digital exploitation after combat becomes as important as searching a weapons cache.

The second response concerns AI platforms. Filters that block an explicit request for an explosive are not sufficient if the fighter breaks the question into several seemingly innocuous pieces. For an operator on the ground, this is not a philosophical question about AI security. It is a question of whether a patrol jumps or returns.

Finally, regional armies will have to accept that the adversary learns quickly. For fifteen years, Boko Haram and ISWAP have survived offensives, airstrikes, internal splits, and victory announcements. Their strength is not only their fanaticism. It is their ability to absorb a useful technique and adapt it to the track, the village, the canoe, or the motorcycle.

In Borno, the operational alert is not just that a chatbot can write propaganda. The alert is that it can help a small group solve a specific tactical problem before the next attack.

The next area to watch is the arc that runs from the Sambisa Forest to the outskirts of Baga and the islands of Lake Chad. If units begin to cross trenches more easily, reduce their losses in ambushes, or standardize new explosive devices, one will need to look beyond the hand of the explosives expert. There may also be a prompt typed on a phone.