Nairobi, Kenya — Artificial intelligence safety research company Anthropic has shut down a Kenyan AI-driven propaganda campaign that was praising the government and attacking the opposition using its flagship AI model, Claude.
The operation involved the creation of multiple fake social media accounts and automated content generation tools. These tools leveraged Claude’s advanced language capabilities to draft persuasive articles, social media posts, and comments designed to sway public opinion and suppress dissenting voices.
Anthropic’s safety team detected the malicious activity through internal monitoring systems designed to flag coordinated inauthentic behavior. The company confirmed it has banned the accounts involved and shut down the operation, emphasizing its commitment to preventing the misuse of its technology.
The shutdown raises urgent questions about the vulnerability of emerging democracies to AI-driven influence operations. As the country approaches its next general election, tech watchdogs are warning that AI tools could be used to amplify misinformation and deepen political divides if left unchecked.
Kenyan digital rights organizations have welcomed Anthropic’s intervention but are calling for stronger local regulatory frameworks to govern the use of generative AI in political campaigning. The incident highlights a growing global trend where bad actors exploit large language models to bypass traditional content moderation.
