Commenting on the launch of the AI Fact-Check Assistant (Aifa) chatbox, Gerakan Media Merdeka’s Radzi Razak warned against an over-reliance on AI chatbots, saying misinformation is often tied to politics, media illiteracy and a lack of access to credible sources.

He said the public tends to believe the responses generated without doing their own verification.

Given the government’s past use of laws to silence critics, Radzi said a government-regulated AI chatbot could be programmed to ignore or downplay certain issues, reinforce official narratives, or even label legitimate criticism as fake news.

“An independent and free media is still the best safeguard against misinformation, not government-controlled AI,” he said.

Communications minister Fahmi Fadzil said Aifa on WhatsApp will help curb the spread of unverified forwarded messages, adding that “fake news spreads faster than Malaysia’s internet speed”.

According to Bernama, users have praised the chatbot for the speed and efficiency with which it verifies information online.

In 2024 alone, the Malaysian Communications and Multimedia Commission (MCMC) requested the removal of 19,546 fake news items from online platforms, with 17,245 content items removed after verification. From 2020 to Jan 27 this year, a total of 25,114 fake news items have been removed.

National Union of Journalists (NUJ) president Low Boon Tat acknowledged the benefit of AI tools but said that they may inherit biases from the data they are trained with.

“Chatbots are trained using data that may have existing inconsistencies or biases, which will naturally flow down to misinformation, instead of correcting it,” he said.

Without proper safeguards, Low said fact-checking could become a tool for censorship rather than for truth-seeking.

“There should be some form of independent oversight for transparency such as an agency that is completely independent of any kind of government influence or a coalition of academia, civil servants and even international organisation,” he said.