The Malaysian Communications and Multimedia Commission has ramped up its battle against manipulated media, successfully removing 12,353 deepfake posts between January and June 2024 out of 13,122 removal requests submitted to social media platform operators. The 94 per cent takedown rate, disclosed in a parliamentary reply by the Ministry of Communications, underscores the scale of artificial intelligence-generated content flooding Malaysian digital spaces and the regulator's aggressive response to the threat.
The figures reveal a staggering volume of deceptive content targeting Malaysians. During the same six-month period, the MCMC also requested removal of 275,787 posts related to scams, fake accounts, and impersonation attempts, with platforms successfully taking down 262,293 items—a 95 per cent success rate. The consistency in these removal percentages suggests that major platform providers, when formally notified, are responsive to Malaysian regulatory authority, though questions remain about how many manipulated posts escape detection before removal requests are issued.
In response to parliamentary questions about artificial intelligence misuse, particularly regarding deepfakes and identity fraud, the government outlined its multi-pronged regulatory approach. The framework combines reactive content removal with proactive labelling requirements designed to inform users about media authenticity before deceptive content spreads widely. This strategy acknowledges a fundamental challenge in the digital age: once false content circulates, removing it afterwards cannot fully undo reputational damage or financial losses incurred by victims.
A cornerstone of Malaysia's enforcement toolkit is the Risk Mitigation Code, which became effective on June 1 this year. The code mandates that licensed platform operators attach labels to content generated or substantially altered using artificial intelligence technologies, including deepfakes and manipulated images or audio. The requirement represents a significant shift in platform responsibility, placing burden on service providers to identify and transparently flag synthetic media rather than relying solely on user reports or regulatory intervention.
Parallel to content removal efforts, authorities have also begun invoking provisions in the Online Safety Act 2025, a newer legislative tool with apparent narrower application. Between January and June, the MCMC submitted only five removal requests under this act, specifically targeting financially motivated scams. All five pieces of content were successfully removed, suggesting that when regulators zero in on specific illegal activities like financial fraud, platform cooperation is absolute and swift.
Criminal enforcement remains a central pillar of Malaysia's response to online deception, though prosecution numbers appear modest relative to content volume. The MCMC investigated 574 cases involving false online content under Section 233 of the Communications and Multimedia Act 1998 stretching back to January 2022. Of these, 23 cases progressed to court prosecution, with 12 concluded and 11 still ongoing. The relatively low prosecution rate hints at investigative bottlenecks, evidential challenges, or prosecutorial resource constraints within Malaysia's judicial system.
Courts have imposed meaningful financial penalties in concluded cases, with total fines reaching RM79,000 across 12 convictions, yielding an average fine of approximately RM6,583 per case. One offender faced six months' imprisonment after refusing to settle their fine, signalling that Malaysian courts view certain online deception cases as warranting custodial sentences. However, whether such penalties deter potential offenders or meaningfully impact organized actors remains unclear, particularly given the anonymity and jurisdictional complexities inherent in digital crimes.
The enforcement pipeline reveals significant administrative activity beyond prosecution. As of June 30, authorities had issued compound settlement offers in 31 cases totalling RM1.22 million, issued 84 warning letters, and maintained 47 cases under active investigation. An additional cluster of cases was classified as requiring no further action, suggesting that investigators sometimes determine alleged content does not breach applicable law or platform policies. This mix of outcomes indicates that regulatory outcomes extend far beyond courtroom convictions, encompassing graduated responses suited to violation severity.
The case of HarakahDaily's Facebook account exemplifies the nuance in regulatory application. Despite prominent discussions about that outlet's content, the MCMC confirmed as of June 30 that no First Information Report had been filed regarding the account, meaning no formal police investigation had been initiated. The ministry, however, signalled readiness to pursue firm action should any content breach laws or platform guidelines, suggesting that the account remains under surveillance pending specific violations.
For Malaysian readers navigating an increasingly complex information landscape, these statistics carry important implications. The sheer volume of manipulated content—over 13,000 deepfake removal requests in six months alone—demonstrates that artificial intelligence-based deception has moved from theoretical concern to practical, pervasive threat. The new labelling requirements may provide some protection by flagging synthetic media, yet educated digital literacy remains crucial since labels themselves can be dismissed or overlooked by vulnerable populations.
Regionally, Malaysia's enforcement approach offers a template for other Southeast Asian nations grappling with deepfakes and scams. The combination of content removal, labelling mandates, and criminal prosecution creates multiple barriers to deceptive actors. However, the enforcement model also raises questions about consistency, appeals rights, and whether removal-based strategies adequately protect citizens before harmful content circulates widely across borders.
Looking forward, the effectiveness of these measures will depend on sustained coordination between platform providers, regulators, law enforcement, and courts. As artificial intelligence tools become more sophisticated and accessible, the volume of deepfakes and manipulated content will likely accelerate, potentially overwhelming current enforcement capacity. Malaysia's proactive stance—establishing labelling rules and pursuing prosecutions—positions the country relatively well, yet the underlying tension remains: reactive enforcement always struggles against rapidly evolving technology.
