Malaysia's fight against artificial intelligence-generated falsehoods is intensifying, with authorities reporting the removal of more than 11,600 instances of deepfake and false content following takedown requests issued to social media companies. Deputy Communications Minister Teo Nie Ching disclosed in Parliament that the Malaysian Communications and Multimedia Commission (MCMC) has processed nearly 12,500 complaints related to AI misuse since the start of 2024, reflecting a concerning trajectory of digital deception that threatens public trust and democratic discourse across the nation.
The scale of the problem has grown dramatically. Deepfake and AI-generated false content complaints jumped from just 917 reports in 2024 to 3,612 cases in 2025, then surged to 7,967 as of mid-June this year. This eightfold escalation underscores how quickly synthetic media technologies are being weaponised for misinformation campaigns, financial fraud, and reputation attacks. The acceleration suggests that awareness of reporting mechanisms is improving, but also that bad actors are becoming more sophisticated and prolific in their deployment of these tools.
For Malaysia, the timing of this enforcement push is strategically important. The nation has been navigating complex questions around digital regulation, balancing legitimate speech protections with security concerns. The introduction of the Online Safety Act 2025 and its accompanying Risk Mitigation Code (RMC) represents a legislative pivot that makes social media platforms—not just individual users—accountable for harmful AI-generated content on their networks. This framework shifts responsibility upstream, requiring licensed platforms to implement active defences against synthetic media abuse rather than merely responding to complaints after harm has occurred.
The MCMC's operational response has become more granular. Beyond simply removing flagged content, the commission now conducts technical assistance for law enforcement agencies, providing forensic analysis and profiler information that helps investigators trace the origins and distribution patterns of deepfakes. This forensic capability is crucial for identifying organised operations versus opportunistic misuse, and potentially tracing funding or coordination with domestic or foreign actors seeking to destabilise Malaysian institutions or manipulate public opinion.
Platform compliance under the new regulatory regime carries teeth. Licensed social media providers must now implement risk mitigation measures specifically targeting AI-generated content. Companies found breaching these obligations face prosecution under the Online Safety Act, with convictions carrying fines of up to RM1 million, plus additional financial penalties reaching RM10 million. For global platforms operating in Malaysia, these penalties represent meaningful exposure that has prompted engagement with MCMC to audit and upgrade their content moderation systems.
Another dimension of this regulatory approach concerns advertiser verification. Platforms must now confirm the identities of advertisers through official channels such as the Companies Commission of Malaysia, a requirement designed to prevent fraudsters from using fake accounts to run scams or disseminate misleading advertisements. This verification step creates an audit trail and makes it riskier for scammers to use mainstream platforms for coordinated fraud operations, though determined threat actors are likely to find workarounds or migrate to less regulated channels.
The deepfake problem carries particular resonance in the Malaysian context. During election cycles and sensitive political moments, synthetic media depicting political figures could be weaponised to manipulate voters or discredit opponents. Unlike traditional false information, which can be fact-checked through research, deepfake videos are inherently more credible-appearing and difficult for ordinary citizens to authenticate. A convincing video of a political figure making inflammatory statements could spread widely before verification is possible, inflicting reputational damage regardless of its falsity. This asymmetry between creation speed and debunking speed favours malicious actors.
The regional dimension also matters. Southeast Asia faces coordinated disinformation campaigns from state and non-state actors, some international in origin. Malaysia, with its significant Chinese diaspora, multicultural composition, and geopolitical location, is a natural target for actors seeking to sow division or manipulate public opinion on sensitive issues. The surge in deepfake complaints may reflect not only increased detection capacity but also genuine escalation in offensive operations targeting the country.
Government technical capacity is improving, though challenges remain. The MCMC's proactive monitoring of social media for AI-generated content requires substantial computational resources and specialist expertise. As AI generation tools become more accessible and sophisticated, the agency faces a moving target—tools that generate convincing synthetic media are becoming easier to use and harder to detect. Staying ahead of this technological curve requires continuous investment in detection algorithms, international collaboration with other nations facing similar threats, and information-sharing with platform companies about emerging synthesis techniques.
Public awareness remains underdeveloped. While authorities are removing deepfakes and prosecuting bad actors, most Malaysians lack practical skills to identify synthetic media. Educational campaigns explaining how deepfakes are created, what telltale signs to look for, and how to verify content before sharing could reduce the amplification of false material. Some platforms have begun adding labels to AI-generated content, but adoption remains uneven and easily circumvented by determined manipulators.
The legal landscape continues evolving. Beyond the Online Safety Act 2025, Malaysia will likely see additional legislation targeting specific harms associated with deepfakes—from election interference to non-consensual synthetic intimate imagery. Other democracies facing similar challenges have moved toward criminal penalties for creating and distributing deepfakes without consent, particularly those depicting real individuals in false compromising situations. Malaysian policymakers will probably monitor international precedents as they refine their approach.
Enforcement effectiveness ultimately depends on sustained resourcing and inter-agency coordination. The MCMC cannot act alone; police cyber units, the Attorney General's Chambers, and content moderation teams at platforms must function as an integrated system. Gaps in this system create opportunities for malicious actors. Malaysia's commitment to removing over 11,600 items of deepfake content signals seriousness, but the eightfold growth in complaints suggests the nation is still in the early stages of a much larger challenge that will shape digital governance for years to come.
