Kenya has been named in a new global report on artificial-intelligence misuse after Anthropic said it disrupted an alleged influence operation that used its Claude AI system to produce political messaging at scale.
According to Anthropic’s threat-intelligence findings, a Kenya-based pro-government operator allegedly used Claude to generate social-media posts praising government figures while also pushing negative narratives about political opponents ahead of the 2027 General Election.
The concern is not simply that political messaging was created with AI. Political campaigns have long used digital teams, influencers and coordinated online communication. The deeper concern is the reported effort to make centrally produced content appear to be authentic grassroots opinion from ordinary citizens.
That blurs the line between public debate and manufactured support.
Anthropic said the operator used AI to draft and adapt posts for multiple accounts, allowing a small team to create far more content than would normally be possible. The system could reportedly help produce messages in different tones, respond to online discussions and target particular political narratives.
The company said it removed the accounts involved after identifying the activity. Its broader report details attempts by different actors across the world to misuse AI for cyber operations, propaganda, surveillance and other harmful activity.
For Kenya, the allegations arrive at a sensitive moment. The country is moving steadily towards the 2027 election, when social-media platforms will again be central to political mobilisation, public argument and the spread of information.
AI can be used responsibly for research, translation, education, customer service and creative work. But its ability to create convincing text, images and videos also makes it easier to scale disinformation and coordinated influence campaigns.
The report should prompt urgent questions for political parties, regulators, technology companies and voters. Who is financing online political campaigns? Are sponsored messages clearly labelled? How can the public distinguish genuine citizen opinion from organised digital operations?
The answer cannot be censorship. Kenya needs transparency, media literacy and clear rules requiring political actors to disclose coordinated and AI-generated campaign material. Platforms must also act faster against fake accounts and deceptive networks.
As 2027 approaches, Kenya’s democratic contest will not be fought only on campaign stages and in rallies. It will also be fought in timelines, comment sections, WhatsApp groups and algorithm-driven feeds. The public deserves to know when “grassroots support” is real and when it may have been engineered.
No comments yet. Be the first to share your thoughts.