Anthropic says it disrupted AI election-manipulation platform targeting Malaysia

US AI firm Anthropic said it disrupted a commercial election-manipulation platform that used its Claude models to target voters across Malaysia, The Edge reported on 11 September. According to Anthropic's threat report, the network comprised roughly 1,000 fake X accounts, a fake news outlet and fabricated dossiers, and the Malaysian Communications and Multimedia Commission (MCMC) denied any involvement.

Key points

• Anthropic said the actor built a voter profiling platform using census and electoral data across all 222 parliamentary constituencies, aimed at race, religion and royalty.
• The firm traced the platform to an Istanbul-based technology company that sold access as a paid service.
• Anthropic said the operator pursued a contract with Malaysia's national communications regulator, but it found no evidence that the effort succeeded.
• MCMC said such activities are not part of its functions or mandate, and that it is examining the report and seeking further information.

Why it matters for Malaysia

The case shows how generative AI could lower the cost of tailored political manipulation in a multiracial society. It also adds weight to Malaysia's wider debate on who should oversee such misuse.

Questions answered

What did Anthropic report about Malaysia?

In a report published on 10 September 2026, Anthropic said it disrupted a commercial election-manipulation platform that used Claude to target Malaysian voters. It described about 1,000 fake X accounts, a fake outlet called Malaysia Pulse and fabricated dossiers, and said all linked accounts were terminated.

Was the Malaysian government linked to the operation?

Anthropic said it found no evidence connecting the operation to the prime minister or his office, and no evidence that the operator's pursuit of a contract with the regulator succeeded. MCMC denied any involvement and said it is examining the report.