A new report reveals that the AI model Claude was allegedly used in pro-Russian AI disinformation campaigns spanning several African nations. The US-based company Anthropic published this detailed account in September.
The report, titled “Detecting and Countering Misuse of AI,” specifically highlighted operations in the Central African Republic (CAR). There, Russian agents purportedly used Claude to bolster pro-Russian influence. This involved monitoring opposition politicians and creating propaganda content.
Claude’s primary use was generating material for Radio Lengo Songo. This radio station, according to the investigative collective All Eyes on Wagner, receives Russian funding. All Eyes on Wagner is an international group monitoring the activities of the Wagner Group and its successor structures, like the Africa Corps.
The AI-generated content consistently promoted a positive image of the Central African government and Africa Corps mercenaries. Conversely, it adopted a critical stance against France and the political opposition. Lou Osborn, from All Eyes on Wagner, contributed to the research.
Osborn noted these methods align with practices previously observed with the Wagner Group in the CAR. He recounted several arrests of individuals not closely tied to President Faustin-Archange Touadera’s regime. These arrests often triggered online and radio campaigns, portraying the individuals as French proxies or coup plotters.
Anthropic further reported requests for Claude to depict opposition figures as armed militants. This aimed to incite security forces against them. Claude reportedly refused these specific requests. However, the report did not name the monitored opposition figures or detail the collected data.
For the opposition, the report underscores the government’s determination, alongside its Russian allies. Martin Ziguele, chairman of the Movement for the Liberation of the Central African People, cited past similar incidents. He mentioned efforts to link opposition leaders to rebel groups following a 2020 attack on Bangui.
Dmitri Sytyi, a representative of the Russian mission in the CAR, denied the allegations. He stated, “These are vague allegations, without precision and without evidence.” Russia firmly refutes the report’s claims.
Nevertheless, the report indicates the CAR was not an isolated case for AI disinformation campaigns. The Democratic Republic of the Congo (DRC), Kenya, Sudan, and Mali were also reportedly affected. In the DRC, Anthropic identified 318 articles on 70 fake news sites, distributed by over 250 fictitious X accounts. This content supported the Congolese government’s stance in its conflict with Rwanda.
In Kenya, Claude allegedly produced batches of 50 messages at a time, supporting the government and criticizing the opposition. Sudan saw AI used in an operation attributed to individuals in the United Arab Emirates. This involved creating fabricated witness statements for the United Nations.
Meanwhile, in Mali, an adviser with intelligence ties reportedly used Claude. This was to develop a platform for monitoring telecommunications. It allegedly targeted 25 million SIM cards, circumventing legal authorization requirements. Anthropic identified and shut down these various operations between December 2025 and August 2026, primarily by deleting accounts and enhancing detection systems.
