PETALING JAYA: Millions of Malaysian voter records were allegedly used to map and target political fault lines across all 222 parliamentary constituencies in an AI-driven influence operation, raising serious questions over how such sensitive data was obtained and used, says Datuk Seri Dr Wee Ka Siong (pic).
The MCA president said Putrajaya must treat the revelation with the highest level of seriousness, warning that the matter was no longer merely about fake accounts, but involved the security of personal data, the integrity of elections and national security.
This follows a report released by Anthropic, the company behind the Claude AI platform, which revealed that Malaysia had been targeted in an AI-driven manipulation operation involving more than 1,000 fake X accounts, voter data and constituency-based targeting systems aimed at influencing political sentiment.
“The people have the right to know the truth. The government cannot remain silent, cannot turn a blind eye and cannot turn a deaf ear.
“A full explanation, a comprehensive investigation and accountability for any parties involved must be provided. Do not defend those who are at fault and do not hide the truth.
“The government must also give a clear assurance to all Malaysians that their data is protected and that manipulation using AI such as this will not happen again.
“If millions of voter records can be used in a political manipulation operation through a foreign AI platform, this is no longer merely an issue of fake accounts, but an issue of the security of the people’s data, the integrity of elections and national security,” he said in a TikTok video.
According to Anthropic’s report, titled “Detecting and Countering Misuse of AI: September 2026”, an alleged operation had used real census data and millions of voter records to target Malaysia’s most sensitive political and social fault lines, including race, religion and the royal institution, across all 222 parliamentary constituencies.
The report said Anthropic exposed the operation after finding that an account had used its platform to manipulate elections and target Malaysian users.
Dr Wee said the revelation was concerning and described the operation as serious political manipulation.
He also questioned the source of the voter records and whether they had been leaked, bought, collected from open sources or obtained from a system that was supposed to be protected.
“This is not basic political manipulation, but serious manipulation using AI. We are not talking about a small system. Claude is among the most advanced generative AI models today, in the same class of technology as ChatGPT.
“What is even more worrying is that this data was said to have been used to exploit the country’s most sensitive issues, namely the 3Rs – race, religion and royalty.
“This is very dangerous to the personal security of the people. It is clear as daylight that the people have been betrayed and our data was leaked to a foreign AI platform for the purpose of manipulating elections.”
On Friday, the Malaysian Communications and Multimedia Commission (MCMC) said it was seeking further information to establish the facts and understand the context of the issues raised.
The commission, however, emphasised that the matters outlined in the report did not fall within its functions or mandate and that it was not involved in any activities or dealings as described in the report.
@drweekasiongData peribadi jutaan rakyat Malaysia bocor dan disalahguna untuk manipulasi pilihan raya? Dari mana data pengundi rakyat Malaysia ini diperolehi? Siapa jual? Siapa beli? Ini merupakan isu serius melibatkan integriti dan keselamatan negara! original sound - DrWeeKaSiong
