Skip to content
Submit incident
AII-1000-53691
ConfirmedA court, a regulator or the organization itself has stated the core facts, or the record cites a court or regulator source.

Anthropic detects AI-generated political propaganda campaign in Kenya

Anthropic shut down an account used by a Kenyan political operator to generate AI-produced social media content designed to appear as grassroots support for government officials ahead of the 2027 election. The operator used Claude across multiple sessions to create batches of 50 tweets praising Energy Cabinet Secretary Opiyo Wandayi, explicitly instructing the AI to disguise the coordinated campaign as spontaneous public sentiment.

11 September 2026 Β· Anthropic Β· Misinformation and synthetic media
11 September 2026
AII-1000-536912 sources ↓
Quick facts
Category
Misinformation and synthetic media
Country
Kenya
Source
Standard Digital News - Kenya
Date
11 September 2026
Drafted by
Raidu Incident Detection

What happened

Anthropic identified and shut down an account operated by a single domestic actor in Kenya who used Claude to mass-produce political content designed to appear as spontaneous grassroots sentiment. The operator generated batches of exactly 50 tweets across several sessions, explicitly instructing Claude to make posts look like grassroots commentary rather than a coordinated campaign.

The campaign praised Energy Cabinet Secretary Opiyo Wandayi for stopping a scheduled Kenya Power electricity tariff hike, using hashtags #PowerReliefKE and #PoweringTheNewKenya. The operation used political scripts, talking points and hashtags, instructing Claude to format posts into deployment-ready content ahead of Kenya's 2027 General Election.

Reported impact

The campaign was designed to manipulate public perception by disguising coordinated political messages as organic grassroots support for government figures ahead of the 2027 election.

Index analysis

Anthropic tracked the operation under the technological designation GTG-54004 and said the actor had instructed the AI model to humanise and format posts into cross-platform interactive widgets.

Governance controls

Anthropic identified and removed the account that was misusing its Claude service for mass-producing propaganda.

Classification

Category
Misinformation and synthetic media
Industry
Media, social platforms and marketing
Country
Kenya
Harm status
Potential harm
Intent
Intentional
Timing
Post-deployment
Harm types
Public interest
Harmed
General public
Failure modes
Manipulation or dark pattern
Consequences
None recorded, System suspension or termination, Content removal
AI systems
Claude

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

Standard Digital News - Kenya
How Claude AI was misused to amplify praise for CS Wandayi
11 September 2026
businesstoday.co.ke
Anthropic detects AI-generated political propaganda campaign in Kenya
Sep 2026AI Chatbots Amplify Misinformation from Partisan "Pink Slime" News SitesConfirmedMisinformation and synthetic mediaPChome Online ζ–°θžYahoo News (Taiwan)δΈ­εœ‹ε ± China Press+5Sep 2026Anthropic disrupts AI platform used for Malaysian election manipulationUnder investigationMisinformation and synthetic mediaThe IndependentBianet - Bagimsiz Iletisim AgiThe Sun Malaysia+6Sep 2026Former Sputnik Moldova editor used Claude to generate disinformation about President SanduConfirmedMisinformation and synthetic mediaHotNews.roRadio Europa LiberΔƒJul 2026Study finds Claude cited Russian disinformation network in news responsesReportedMisinformation and synthetic mediaBildGMXMay 2026CCP state media content found in AI chatbot training dataConfirmedMisinformation and synthetic mediaThe Epoch Timesepochtimes.com