AI deepfakes and fake fighters reshape Mali’s information war

Relevance to the sub: Use of AI tools in the information battlefield of the war in Sahel, allowing for the expansion of existing efforts and new approaches to psychological warfare and recruitment tactics

The Dakar-based Timbuktu Institute, which studies political, cultural and military developments in Western Africa, has released a report – direct link here (pdf, in French) – examining how factions warring for the control of Mali have adopted AI tools to reshape their propaganda campaigns, turning Mali into a "laboratory of AI-assisted informational warfare".

The report identifies a more "basic" use of AI by JNIM, the Al-Qaeda affiliated Islamist insurgents, who are using AI tools to cut down on production costs for their communication efforts, and notably by using auto-translation features to translate their propaganda in the many languages spoken throughout the country to reach more communities – especially the Bambara, Fulani and Songhay populations who could provide the group with potential recruits. The productivity gain allows the group, who is also taking inspiration from the Islamic State's previous use of AI, to significantly increase the volume of their content to flood the zone.

The FLA, a Tuareg separatist group that concluded an alliance with JNIM against the Malian government, uses AI in a more decentralized manner, and has specialized in the creation of AI-generated "characters" whose diatribes against the military regime are reposted and amplified by complex networks of bots on social media platforms, notably by artificially boosting the number of followers and "likes" to create the illusion of a much greater popular support for the separatists' cause.

But the most complex and advanced use of AI is found among the Africa Corps, the Kremlin-managed paramilitary created on the foundations of Wagner, whose founder Yevgeny Prigozhin had pioneered mass use of troll farms and online disinformation.

Russia's online informational warfare in Africa is a longtime practice that has deepened with Russia's direct military involvement in Mali, Burkina Faso and Niger. As early as 2023, reports from NGOs and intelligence services evidenced that Russian operatives had planted deepfakes as part of their informational warfare to "prepare the field" before they deployed fighters to Burkina Faso, and attempted to divert OpenAI's early functions in 2024 to generate en masse fake profiles and comments on West African social media spaces glorifying Russia and lambasting France and the West for their past interventions.

According to the report, the Malian military regime has fully integrated Russian disinformation tools to generate and amplify narratives accusing France, ECOWAS or the UN of attempting to destabilize the country, but also to discredit and harass domestic opponents, lawyers, journalists or NGO workers who decry the military regime's crackdown on human rights. Combined with the junta's lockdown of the national information space by their tight control of the media and their persecution of independent journalists, the use of AI-assisted disinformation seems to be effective in strengthening the junta's grip on the country.

In conclusion, the mass adoption of AI tools is acting as a force multiplier for existing disinformation practices, and will continue to be integrated by combattants on all sides to boost their propaganda efforts as well as their approach of recon and drone warfare, another key innovation of 2020s warfare.