Kenyan political operator used Claude to mass-produce election propaganda, Anthropic says

Anthropic said a Kenyan political operator used the company's Claude chatbot to mass-produce social media posts praising government officials and targeting opposition figures ahead of the 2027 general election, per a report released on September 11. TechCabal reported that the operation did not reach or influence real people, but Anthropic classified it as evidence of how generative AI could shape Kenya's political ecosystem by enabling actors to manufacture coordinated propaganda that appears organic.

Anthropic identified and removed an account used by a single actor to generate Kenyan political content in batches of exactly 50 tweets. The actor explicitly instructed Claude to make the posts appear to be spontaneous grassroots commentary rather than part of a coordinated campaign. The content praised Energy Cabinet Secretary Opiyo Wandayi over a Kenya Power tariff issue, used hashtags such as #PowerReliefKE and #PoweringTheNewKenya, and pushed claims that the United Opposition coalition was breaking apart ahead of the election. Other posts targeted former deputy president Rigathi Gachagua and former president Uhuru Kenyatta. Anthropic said the campaign used a single playbook to amplify pro-government and anti-opposition narratives, a pattern consistent with a coordinated electoral communications effort.

Anthropic found no evidence of government involvement and stopped short of linking the operation to the Kenyan government or any political party. The company described the activity as consistent with "an entirely domestic Kenyan operation" and said the network's pro-administration tone and specific hashtags suggested it was "plausibly aligned with the ruling coalition," though it could not identify the organization or people behind it. The same actor used an almost identical workflow to market Kenyan retail brands under a separate marketing persona, supplying political topics, talking points, and hashtags before using Claude to turn them into large batches of platform-ready posts. Anthropic emphasized that "the core ideological messages were completely decided by the actors before Claude was used," meaning AI helped produce large volumes of content but did not generate the underlying political strategy.

Anthropic classified the Kenyan operation as Category One on the Breakout Scale, meaning the activity remained contained within its own network of fake accounts and local influencers on a single platform with no evidence that the campaign reached or influenced real people. The disclosure adds to a growing pattern of AI-enabled political interference: OpenAI disrupted election-related AI activity in Ghana in January 2025, and Anthropic's report also identified an influence operation targeting audiences in the Democratic Republic of Congo.