Social media platforms’ non-compliance with AI deepfake fraud removal requests prompts early enforcement of Online Safety Act 2025
Malaysian Communications and Multimedia Minister Fahmi Fadzil strongly criticized some social media platforms for ignoring AI-based fraud content takedown orders and declared the government’s intention to enforce stricter regulations through the early implementation of the Online Safety Act 2025. This decisive action comes as AI-generated deepfake investment scams have reached alarming levels, victimizing even the Yang di-Pertuan Agong (King) and the Prime Minister, establishing what could become a new regulatory benchmark for AI content governance across the Asian region.
Regulatory Impact and Compliance Requirements
The Malaysian Communications and Multimedia Commission (MCMC) submitted 1,602 requests for AI-based scam content removal between January 1 and August 31 this year, with only 1,346 successfully taken down, reflecting an 84% compliance rate. This means approximately 256 pieces of AI fraud content remain active online, representing a 16% non-compliance rate that the government deems unacceptable. The emergence of deepfake investment scams impersonating the nation’s highest leaders has exposed critical gaps in current self-regulatory frameworks. In response, the government has decided to enforce the Online Safety Act 2025 by the end of this year or early 2026, mandating AI content labeling requirements for all social media and online messaging platforms operating within Malaysian jurisdiction.
Mandatory Compliance Standards and Obligations
Under the new Online Safety Act 2025, all platform operators must clearly label AI-generated or AI-enhanced images and content with markers such as “AI generated” or “AI enhanced.” Service providers will also be required to submit comprehensive online safety plans as proof of regulatory compliance, while MCMC is considering imposing licensing requirements on applications utilizing AI technology. These compliance standards will apply universally to all online platforms serving Malaysian users, regardless of platform size or user base. Non-compliance is expected to result in severe penalties including service blocking and substantial fines, reflecting the government’s zero-tolerance approach to AI-enabled fraud.
Industry Response and Adaptation Strategies
Global social media platforms have historically shown reactive approaches to AI content safety issues, but Malaysia’s firm regulatory stance is likely to force comprehensive strategy reassessments. The government’s determination that an 84% compliance rate remains insufficient signals the need for platforms to significantly enhance their AI content detection and removal systems. To improve digital crime response effectiveness, the government has established the National Scam Response Centre (NSRC), led by police in collaboration with the communications ministry, home ministry, Bank Negara Malaysia, and MCMC, creating a rapid response ecosystem for digital fraud cases.
International Trends and Global Standardization
Malaysia’s mandatory AI content labeling policy positions the country alongside the EU’s AI Act and China’s deepfake regulations as a leader in AI governance frameworks. As the first Southeast Asian nation to establish concrete regulatory legislation for AI content, Malaysia’s approach is expected to influence similar policy adoptions across Indonesia, Thailand, Vietnam, and other regional neighbors. Global platforms now face the increasing complexity of simultaneously complying with diverse AI content regulations across different jurisdictions, intensifying the need for international standardization discussions and harmonized regulatory approaches.
