What if the government praise you see flooding social media is not coming from Kenyans at all? Anthropic has exposed a Kenyan political influence operation in which its Claude AI chatbot was allegedly misused to manufacture social-media posts that made government support appear to be genuine grassroots opinion while attacking opposition figures.
Artificial intelligence is emerging as a new weapon in Kenya’s political communication after Anthropic revealed that its Claude platform was used to produce coordinated political content ahead of the 2027 General Election.
According to Anthropic’s September 2026 Threat Intelligence Report, a single actor used Claude to mass-produce Kenyan political posts designed to look like genuine comments from ordinary citizens. The operator repeatedly instructed the AI to generate batches of exactly 50 posts, with the content crafted to appear spontaneous.
The campaign reportedly promoted government-aligned narratives while attacking the opposition. It also pushed messages around issues such as electricity tariffs and narratives portraying the opposition as divided ahead of the 2027 elections.
Anthropic said it found no evidence that the Kenyan government was directly involved in the operation, describing it instead as a domestic political astroturfing campaign designed to manufacture the appearance of widespread public support.
The revelations have raised fresh concerns about the credibility of political conversations taking place online, particularly as Kenya approaches the 2027 General Election.
With AI capable of producing convincing political messages within seconds, the distinction between genuine public opinion and manufactured online support could become increasingly difficult for voters to establish.
The controversy now goes beyond Claude: if political actors can use AI to manufacture thousands of seemingly authentic voices, how will Kenyans know which opinions online are real?
