From the journal

Malaysia National AI Office Opens Consultation on Proposed AI Governance Bill, 10 July 2026

On 10 July 2026, Malaysia's National Artificial Intelligence Office (NAIO) published a Public Consultation Paper on the proposed AI Governance Bill. The Bill would establish Malaysia's first statutory AI framework, adopting a risk-based model and covering the full AI lifecycle from design through retirement. It includes specific provisions on deepfakes, AI-generated copyright, and data sovereignty.

3 min read

Malaysia's National Artificial Intelligence Office (NAIO) released a Public Consultation Paper on 10 July 2026, seeking input on the proposed structure of a dedicated AI Governance Bill. The Bill, when enacted, will constitute Malaysia's first statutory framework specifically governing artificial intelligence systems across all economic sectors. The NAIO has stated the Bill is expected to be presented to Parliament before end-2026, following Cabinet presentation as the next procedural step.

The Consultation Paper proposes a risk-based classification model in which AI systems are assigned to risk tiers with proportionate obligations at each tier. High-risk AI applications would require mandatory AI impact assessments before deployment, incident reporting within defined timelines, and ongoing compliance with baseline operational safety standards. All AI systems regardless of risk tier would be subject to minimum ethical and safety principles. Three specific statutory provisions are proposed: deepfake offence provisions addressing AI-generated synthetic media used to cause harm; copyright attribution rules for AI-generated content and liability for AI training data use; and data sovereignty requirements restricting cross-border transfer of AI training datasets containing personal data of Malaysian persons.

The Bill, in the form proposed, will affect technology companies developing AI products for the Malaysian market, financial institutions and licensed payment service providers using AI in credit decisions and fraud detection, healthcare operators deploying AI diagnostics or triage tools, and digital platform operators running AI recommendation or content moderation systems. Foreign operators delivering AI-enabled services to Malaysian users without local establishment may fall within scope where the AI system's outputs are directed at Malaysian persons, following the extraterritorial approach in the Personal Data Protection Act 2010 and the EU AI Act.

The Consultation Paper is a pre-legislative document and the final Bill may differ materially. The NAIO has not published a deadline for consultation responses. Open questions include the precise definition of high-risk AI applications in the Malaysian context, which sector regulators will enforce the obligations in their respective domains, the treatment of open-source AI models and small-scale operators, and whether the deepfake offence provisions will require specific intent or apply on a strict liability basis.

Licentium monitors APAC AI regulatory developments and advises multinational operators entering or operating in Southeast Asian markets. We assist with AI governance bill consultation submissions, AI governance programme design calibrated to APAC requirements, and comparative analysis against the EU AI Act and other binding frameworks. Work we undertake includes APAC AI regulatory monitoring, AI Governance Bill consultation response, market-entry compliance review, and AI risk classification mapping.

Source: Malaysia National Artificial Intelligence Office (NAIO), AI Governance Consultation Paper, 10 July 2026

AI Regulatory

More from the journal

See all
Illia Prokopiev

EU AI Act Deadlines, Data-Lineage Obligations, Regulated Use Cases, And Penalties

The analysis concerns high-risk AI systems under Article 6. Separate obligations concerning general-purpose AI models are outside scope except where a third-party model is incorporated into a high-risk system.

FTC Characterises AI Output Steering as Deception Under FTC Act Section 5, July 2026

On 7 July 2026, the Federal Trade Commission published proposed policy statement 2026-13628 in the Federal Register. The statement identifies AI output steering - deliberate modification of a system's outputs to advance undisclosed objectives - as potentially deceptive under Section 5 of the FTC Act, 15 U.S.C. § 45. The FTC further concludes that state laws compelling suppression of accurate AI outputs, including Colorado's Artificial Intelligence Act, may be impliedly preempted. Public comment closes 31 July 2026.

European Commission Finalises Article 50 AI Transparency Code of Practice, Effective 2 August 2026

On 10 June 2026, the European Commission published the final Code of Practice on Transparency of AI-Generated Content, setting out technical specifications through which providers and deployers of generative AI systems can demonstrate compliance with Article 50 of EU Regulation 2024/1689. The transparency obligations in Article 50 enter into application on 2 August 2026. Organisations seeking to be listed as initial signatories must submit adherence forms by 22 July 2026.