The Malaysian Communications and Multimedia Commission has intensified its battle against synthetic media and online manipulation, revealing that it successfully convinced social platforms to delete more than 12,000 deepfake-related posts during the opening six months of this year. Authorities submitted 13,122 removal requests to licensed service providers between January and June, with platforms complying in 94 per cent of cases—a notably high success rate that underscores both the clarity of Malaysia's regulatory framework and the willingness of major technology companies to cooperate with national authorities.
The scale of this enforcement effort reflects growing alarm across Southeast Asia about the weaponisation of artificial intelligence for malicious purposes. Deepfake technology, which uses machine learning to create convincing but fabricated video and audio content, poses particular risks during election cycles and periods of political sensitivity. Malaysia's achievement in removing 12,353 posts demonstrates that technological solutions exist when regulators have sufficient legal authority and platform cooperation, yet the sheer volume also illustrates how quickly synthetic content proliferates across social networks before removal.
Beyond deepfakes, the MCMC's broader content moderation efforts reveal an even starker picture of online manipulation across Malaysian digital spaces. Between the beginning of 2022 and June 2024, the commission requested removal of 275,787 pieces of scam-related content, including fake accounts designed to impersonate legitimate users and organisations. The 95 per cent removal rate for these cases suggests platforms have standardised procedures for identifying and taking down mass-produced fraudulent material, though the absolute numbers indicate scams remain a persistent problem requiring sustained attention.
A significant regulatory development took effect on June 1 when the Risk Mitigation Code came into force, establishing mandatory requirements for licensed platform providers to apply visible labels to any content generated or altered using artificial intelligence systems. This marking requirement addresses a fundamental challenge in combating deepfakes: users often cannot distinguish authentic material from synthetic alternatives without explicit disclosure. By requiring platforms to flag AI-created or AI-manipulated content, Malaysia is attempting to build digital literacy and trust at the source, though effectiveness will depend on consistent implementation across all major social networks operating within Malaysian jurisdiction.
The regulatory architecture protecting Malaysian users has expanded with the introduction of the Online Safety Act 2025, which provides fresh enforcement mechanisms specifically targeting financial crimes committed through digital channels. During the first half of this year, authorities submitted five removal requests under this new legislation addressing financial scams, and all targeted content was successfully removed. While the number appears modest compared to broader enforcement figures, it reflects the Act's recent implementation and suggests this tool will become increasingly important as authorities identify patterns of organised digital fraud.
Prosecution activity demonstrates that the government is willing to pursue criminal charges against serious offenders rather than relying solely on administrative removal. The MCMC investigated 574 cases involving false online content under the Communications and Multimedia Act 1998 between January 2022 and June 2024, though prosecution rates remain relatively low. Only 23 cases proceeded to court, with 12 concluding in conviction and 11 still pending. The 12 concluded cases resulted in fines totalling RM79,000, while a single offender was imprisoned for six months after refusing to pay a fine, indicating courts view persistent non-compliance seriously.
Alternative enforcement mechanisms have absorbed the majority of investigated cases. As of June 30, authorities had offered compounds—financial settlements without formal court proceedings—in 31 cases involving a total of RM1.22 million, issued 84 warning letters to individuals or entities who violated standards but did not warrant prosecution, and maintained 47 cases under ongoing investigation. A substantial number of cases were classified as requiring no further action after initial review determined the content fell outside regulatory jurisdiction or did not breach established laws and platform guidelines. This tiered approach allows regulators to deploy proportionate responses while reserving limited court resources for the most egregious violations.
The case of HarakahDaily's Facebook account illustrates the challenges regulators face when determining whether content violates law or platform standards. Although the MCMC stated it had not received a formal complaint by June 30, it signalled readiness to take action if content was found to breach applicable regulations or social media platform terms of service. This proactive stance reflects recognition that political discourse platforms sometimes operate in grey zones where the line between legitimate speech and regulatory violation remains contested and context-dependent.
For Malaysian citizens and businesses, these enforcement statistics carry important practical implications. The high removal rates suggest that reporting deepfakes, scams, and fraudulent content to platforms through standard mechanisms tends to succeed, encouraging victims and concerned users to participate actively in content moderation rather than assuming reports disappear into bureaucratic black holes. Simultaneously, the mandatory AI labelling requirements mean users encountering political videos, celebrity endorsements, or other potentially sensitive content should increasingly see clear indicators when artificial intelligence played a role in creation or alteration.
Regionally, Malaysia's enforcement approach offers both a model and cautionary lessons for other Southeast Asian democracies grappling with synthetic media and online fraud. The combination of legislative authority, regulatory capacity, platform cooperation, and graduated enforcement—from removal to warnings to prosecution—represents a comprehensive strategy. However, the persistence of such content despite high removal rates suggests that supply-side enforcement alone cannot solve the problem; demand-side solutions including digital literacy, media verification skills, and critical consumption habits remain essential complements to regulatory action.
The government's parliamentary disclosures demonstrate transparency about enforcement efforts, though gaps remain in public understanding of how decisions are made about what constitutes illegal deepfakes versus permissible satire or artistic expression. As artificial intelligence capabilities continue advancing and becoming more accessible, Malaysia's regulatory framework will face increasing pressure to balance innovation protection with public safety—a tension particularly acute in a region where misinformation has demonstrable real-world consequences for communal harmony and electoral integrity.
