Anthropic disrupts AI platform used for Malaysian election manipulation
Anthropic reported discovering and disrupting a commercial election manipulation platform that used Claude to deploy fake social media accounts, fabricated news, and voter profiling tools targeting Malaysian constituencies. The Malaysian Communications and Multimedia Commission has opened an investigation into the operation.
What happened
Anthropic reported disrupting a commercial election manipulation platform that used Claude to target Malaysian voters. The operation deployed approximately 1,000 fake X accounts managed through the platform with mechanisms to change cookies and IP addresses to evade detection, a fabricated news outlet, and false intelligence dossiers.
The platform covered all 222 parliamentary constituencies and used census and electoral data to identify and target voters by race, religion, and the monarchy. According to Anthropic's threat documentation review, the infrastructure was marketed as a 'military-grade, AI-driven, real-time political operations ecosystem.'
Reported impact
MCA Youth and Barisan Nasional stated the alleged AI-powered influence operation posed a threat to Malaysia's electoral integrity ahead of the 16th General Election. The operation enabled mass fabrication of identities, narratives and manufactured public sentiment at a previously unattainable scale.
Index analysis
The operation was identified and disrupted by Anthropic's Threat Intelligence team following observation of suspicious Claude usage patterns over eight months. Malaysia's Communications and Multimedia Commission began reviewing the matter and sought clarifications on the scope and origins of the operation.
Governance controls
MCA Youth called for a joint investigation by the Malaysian Communications and Multimedia Commission, CyberSecurity Malaysia and police to identify the domestic or foreign backers without fear or favour. The opposition also called for comprehensive legislation and enforcement mechanisms to address the misuse of generative AI for deepfakes, astroturfing and coordinated disinformation.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.