Artificial intelligence is already being used to manufacture apparently grassroots political support ahead of Kenya’s 2027 election, according to a new threat report by AI company Anthropic, offering an early warning for Nigeria and other African countries preparing for elections next year.
Anthropic said it removed an account used by a single actor to mass-produce Kenyan political content with its Claude AI. The operation generated batches of exactly 50 posts designed to look like spontaneous public commentary rather than a coordinated campaign.
The network praised Kenya’s Energy Cabinet Secretary Opiyo Wandayi over a proposed electricity tariff hike and promoted the hashtags #PowerReliefKE and #PoweringTheNewKenya. It also circulated claims that the United Opposition coalition was breaking apart, targeting politicians including Rigathi Gachagua and former President Uhuru Kenyatta.
Anthropic said it found no evidence of government involvement and considered the activity consistent with a domestic Kenyan political operation. The company said the operation appeared “plausibly aligned with the ruling coalition”, but it could not identify the organisation behind it.
The operation was classified as Category One on Anthropic’s Breakout Scale because its activity remained isolated within a network of fake accounts and did not reach or influence real people.
But its significance for Africa’s next election cycle lies in how easily AI allowed a small operation to manufacture the appearance of public sentiment.
The same workflow was used for commercial marketing, with Claude turning pre-drafted topics and talking points into batches of posts. Anthropic said the model’s main value to the operation was “volume and the appearance of authenticity” rather than developing the political messages themselves.
That risk is particularly relevant to Nigeria, which, like Kenya, is heading for a general election in 2027.
A July report by SBM Intelligence found that 66.5% of Nigerians surveyed were extremely or very concerned that AI-generated political misinformation could undermine the credibility of the 2027 election. The study, based on 829 respondents across eight states and six geopolitical zones, also found that social media was the primary news source for 52.1% of respondents.
Biological weapons threat
Anthropic’s report, covering activity disrupted between December 2025 and August 2026, shows that election manipulation is only one part of a much broader expansion in AI misuse.
The company identified a network of about 70 fabricated news outlets targeting audiences in Africa and other regions. Some used fake journalists, AI-generated profile pictures and coordinated social-media accounts to make the sites appear to be independent local news organisations. One X account posed as a Congolese civilian “digital army”. Anthropic said it found no evidence that any government directed that activity.
The report also details an AI-assisted surveillance operation involving Mali, as well as other cases in which Claude was used in cyber operations, scams and fraud.
More alarming were five cases involving biological research. Anthropic said users sought assistance with work that could potentially contribute to biological-weapons development, including research involving chikungunya and attempts to increase the ability of avian influenza to infect mammals.
The company did not identify the researchers, institutions or countries involved in those cases. It also cautioned that the research should not automatically be interpreted as an attempt to develop biological weapons.
Biological misuse, it said, is “one of the most serious risks of frontier AI model”. Without the correct safeguards, such capabilities “could have catastrophic consequences”, Anthropic said.
“The same information that can be used to develop a biological weapon could also be used to develop, for example, a vaccine or a cure for a disease,” Anthropic said.
Elsewhere, Anthropic said AI was used in cyber operations against organisations worldwide and in work related to conventional weapons, including missiles and drones.
The cases illustrate a shift in the potential threat posed by increasingly capable AI systems. Rather than simply generating harmful information, the models can assist users with multiple stages of complex operations, allowing them to work faster and at greater scale.
Anthropic said it removed the accounts involved, strengthened its detection systems and shared intelligence with authorities and industry partners where appropriate.
For Africa’s 2027 elections, the Kenyan case offers a more subtle warning than the familiar deepfake of a political leader.
A campaign may not need to fabricate a candidate’s voice or face. It can manufacture the appearance that ordinary people are saying the same thing – at scale, repeatedly and across the platforms where voters increasingly get their news.
Discover more from Pluboard
Subscribe to get the latest posts sent to your email.