An İstanbul-based technology company was linked to an AI-assisted influence operation targeting Malaysian voters and supporting the country’s sitting prime minister, according to a security report released yesterday by the Anthropic AI company.
The company detailed the operation in its report, “Detecting and countering misuse of AI,” which examined what the company described as malicious uses of its AI systems over the past eight months and said it had disrupted.
The alleged ranges from an Iran-backed Houthi group's use of Claude to develop ballistic missile guidance software to cyberattacks, AI supply-chain schemes aimed at stealing tokens and API keys, and influence operations, according to the report.
The company said it identified and removed an account that used Claude to operate a commercial election-manipulation platform primarily targeting Malaysia.
'A political operations ecosystem'
The platform was linked to İstanbul-based BBS Bilişim Teknolojileri, which Anthropic said sold access to it as a paid influence service. The operator’s own documentation described the system as a “military-grade, AI-driven, real-time political operations ecosystem.”
According to the report, the platform used census and electoral data, including millions of voter records, to profile voters across Malaysia’s 222 parliamentary constituencies. Targeting focused on political and social divisions involving race, religion and royalty.
The operation managed more than 1,000 fake accounts on X, using techniques intended to make them appear authentic and evade detection. A dashboard allowed operators to set artificial engagement targets. Anthropic said it observed a request for one million artificial views in support of the sitting Malaysian prime minister.
Malaysia's next general elections are scheduled to take place no later than Feb 2028.
Creating fabricated news
The network also operated a fake news outlet called “Malaysia Pulse.” Claude was used to rewrite legitimate Malaysian reporting and republish it under fabricated bylines. The outlet also republished material from Russian and Chinese state-aligned media while removing references to their state affiliations.
The operators also produced fabricated intelligence dossiers containing false allegations against an opposition politician and civil-society organizations.
Anthropic classified the campaign as "Category Two" on its Breakout Scale, meaning its assets appeared across multiple platforms but showed no evidence of reaching authentic communities. (VK)
