Anthropic Exposes AI-Driven Political Campaign Praising Opiyo Wandayi

Kenya’s Energy and Petroleum Cabinet Secretary, Opiyo Wandayi, has denied any involvement by himself or his ministry in an alleged AI-driven campaign that promoted his handling of electricity tariffs.

Wandayi’s comments followed the release of a September 2026 threat intelligence report by US technology company Anthropic, which identified a coordinated digital influence operation in Kenya. Anthropic said one actor used its Claude AI platform to produce batches of 50 social media posts at a time. 

The posts were designed to appear as spontaneous grassroots comments and frequently praised Wandayi for his role in stopping a planned electricity tariff increase by Kenya Power. The campaign used hashtags including #PowerReliefKE and #PoweringTheNewKenya. Anthropic said, however, that it had found no evidence connecting Wandayi or his ministry to the operation.

“My attention has been drawn to some unverified report alleging the creation of some activity on social media… What matters is evidence,” Wandayi said. 

He stressed that anonymous activity online does not, by itself, establish government involvement. Anthropic classified the operation as “Category One” under its Breakout Scale. This means the activity was limited to a single platform and did not succeed in influencing real audiences.

The findings nevertheless demonstrate how AI can be used to produce large volumes of content intended to simulate public opinion. The technology allowed the actor to turn political talking points into numerous posts that appeared to reflect genuine public support for government positions.

Wandayi also used the hashtag #PoweringTheNewKenya in his public response, the same phrase identified in Anthropic’s report. However, the use of the hashtag alone does not establish a connection between the Cabinet Secretary and the accounts involved in the campaign.

Anthropic said the operation also targeted Kenya’s political opposition. The same network circulated claims about divisions within the United Opposition coalition ahead of the 2027 General Election, including claims involving Rigathi Gachagua and retired President Uhuru Kenyatta.

The activity therefore combined messages supporting government positions with content intended to portray divisions within the opposition. Anthropic’s findings indicate that the network was involved in a wider political communications effort.

The AI system was also used for commercial marketing through personas including “SHANKI” and “Elkins Marketer”. Anthropic said this showed how techniques developed for influence operations are increasingly being applied to commercial promotion.

The company also reported growing use of fabricated journalists, fake spokespeople and impersonated institutions. Such methods can make misleading content appear to come from established and trusted sources.

Kenya’s digital environment has previously been identified as vulnerable to online manipulation. A recent report by the Kenya ICT Action Network (KICTANet) documented fabricated media content and influencer-led narratives portraying the opposition as disorganised.

Analysts have warned that repeated exposure to such content can reinforce echo chambers and increase political polarisation, particularly as political communication increasingly takes place on platforms where engagement and individual personalities can have a greater influence than traditional journalism.