ISIL orders supporters to master AI as a ‘religious obligation’ for global operations

Islamic State militants are being urged to integrate artificial intelligence and conversational chatbots into their operations to bolster the efficacy of their global strike capabilities.

The latest two issues of Voice of Khorasan, the English-language flagship publication of the Islamic State Khorasan Province (ISKP)—the group’s Afghanistan-based affiliate—feature dedicated sections instructing supporters on how to become “responsible mujahideen” through the strategic application of AI.

ISKP has been linked to a string of high-profile atrocities across Russia, Iran, Afghanistan, and Pakistan, as well as the thwarted 2024 plot targeting Taylor Swift concerts in Vienna. Last year, the European Union added the magazine’s publisher, the Al Azaim Media Foundation, to its formal sanctions list.

In the magazine’s most recent dispatch, a segment titled “Ways to use AI chatbots responsibly” frames the technology through a dualistic lens. “AI is like fire. You can use it to light up the house or to burn it down,” the text notes, providing a blueprint for leveraging these tools in religious campaigns and the dissemination of sermons.

“AI is everywhere. Get to know it before it knows you too well,” the publication warns, adding a directive for followers to “raise children and students to be cyber-aware and spiritually grounded.”

The editorial further underscores the utility of AI for conducting anonymous private research, emphasizing its ability to shield users from “sharing sensitive information” and “exposing ourselves to unnecessary risks.”

However, the guide prescribes strict operational security, warning that chatbots should not be used to share personal data that could be “traced back to us,” upload classified files, perform inquiries that are politically or security-sensitive, or seek formal religious rulings.

A preceding issue of the magazine featured a comparative analysis of AI models developed by OpenAI, Anthropic, Meta, Microsoft, and DeepSeek.

The analysis included warnings regarding models allegedly involved in “Israel Defense Forces AI infrastructure” and recommended the privacy-centric Brave Leo tool for “sensitive queries.”

The group’s pivot toward emerging technology coincides with warnings from Jonathan Hall KC, the UK’s Independent Reviewer of Terrorism Legislation. Hall has identified AI as a “new incoming wave” that will amplify terrorist propaganda and facilitate the planning of kinetic attacks.

“Developments in AI and Agentic AI are moving at pace, and more and more of these capabilities are becoming available to terrorist groups,” Hall told POLITICO after reviewing the latest ISKP directives. “LLMs [large language models] lower the barriers to creating sophisticated and bespoke propaganda, and this becomes a problem as terrorist groups increasingly look to exploit local grievances in local languages.”

Hall added: “I would not be surprised to see chatbot radicalization starting to take off: if you can build a terrorist website, why wouldn’t you build a terrorist chatbot? If government officials aren’t following AI developments, they really should be.”

An earlier edition of Voice of Khorasan went as far as to argue that AI literacy has become “fard al-ayn”—an individual religious obligation for every Muslim, comparable to prayer, fasting, or the Hajj pilgrimage.

One note explicitly states: “AI is no longer optional; it is your shield and your compass in a digital world full of hidden threats.”

Avi Jager, Senior Director of the Serious Harms Intelligence Department at AI safety firm Alice, observed that these documents signal more than just “tactical guidance.” They represent an “ideological shift with significant long-term implications,” Jager noted.

“For years, groups like ISIS and the broader jihadi movement were cautious about new technologies like AI. There were even arguments in earlier literature that using such tools could be un-Islamic or that relying on them contradicted the purity of divine guidance,” Jager said.

“That is what makes this moment so remarkable: ISKP is no longer just tolerating AI; it is explicitly endorsing its use under certain conditions.”