Artificial intelligence is rapidly becoming 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 users time can also be misused, and a new report suggests that terrorist organizations are already finding ways to do so.

According to a research paper shared by researchers with The New York Times before publication, researchers found evidence that members of the Boko Haram group were using popular AI-powered chatbots to support daily activities and combat-related tasks.

Interviews conducted with 27 former members in Nigeria over the past two years indicate that tools such as ChatGPT, Gemini, Claude, Grok, DeepSeek, and Meta AI were used to gather technical information, troubleshoot and repair weapons, and even assist in planning attacks, according to a report by Digital Trends, a technology news website, as seen by Al Arabiya Business.

What makes these findings particularly alarming is that this was not described as just isolated experiments by a few individuals using AI. Instead, the newspaper reported that the group's use of AI has become organized, with dedicated teams, internal training, and knowledge sharing among members.

The researchers said that some users managed to bypass built-in safeguards designed to prevent AI from responding to requests related to violence.

This does not necessarily mean that today's AI chatbots provide dangerous instructions on demand; much of the research focuses on activities up to the end of 2024, and AI companies say newer versions of their models include stronger safeguards.

They also indicate that many requests fall into a gray area. For example, asking the AI how to repair a motorcycle or understand basic chemistry is not harmful in itself, even though similar information could be misused.

OpenAI, the developer of ChatGPT, told The New York Times that using its chatbot to support terrorism or violence violates its policies, stressing that it continues to improve safeguards to prevent misuse.

Meta, developer of Meta AI, noted that the research primarily covered older versions of its AI models, adding that it has since strengthened safety measures. In contrast, xAI and DeepSeek did not respond to the newspaper's requests for comment.

Security experts remain cautious about assuming AI suddenly gives terrorist groups superhuman capabilities, as planning and executing attacks still rely on real-world factors such as logistics, funding, communications, and human coordination, which a chatbot cannot replace.

Instead, researchers believe AI may make less experienced members somewhat more capable, rather than radically changing the threat landscape.

However, this research is a new reminder that the debate over AI safety is not limited to students cheating on assignments or creating fake images. As AI models become more capable and accessible, companies will continue to face greater pressure to ensure these powerful tools remain useful for legitimate users while becoming harder to exploit by malicious actors.

Advertising material

Advertising material