The Malaysian government has moved to clarify its position on a contentious issue facing the global creative industries: whether artificial intelligence firms can freely use copyrighted material for training their models. Deputy Domestic Trade and Cost of Living Minister Datuk Dr Fuziah Salleh confirmed during parliamentary proceedings in Kuala Lumpur that copyright owners must grant explicit approval before their works can be utilised in AI development, establishing a clear legal boundary that puts Malaysia broadly in line with stricter international approaches to creator rights.
Fuziah's remarks came in response to pressing questions from members of the Dewan Negara concerning the government's stance on intellectual property protection in the emerging AI landscape. Senator Datuk Wu Him Ven sought clarity on whether Malaysia would require such permissions and what mechanisms might exist to compensate local creators through licensing and royalty frameworks. The deputy minister's response affirmed that existing law already provides these protections without the need for new legislation, at least in principle. The Copyright Act 1987 remains the governing framework, and it does not distinguish between traditional commercial uses and AI-related applications when determining whether permission is required.
According to Section 27 of the Copyright Act 1987, the practical arrangements for compensation remain flexible and subject to negotiation between interested parties. This could take several forms: direct licensing agreements struck between AI companies and individual copyright holders, assignments of rights that transfer control to the technology firms, or arrangements administered through Collective Management Organisations that represent groups of creators. The deputy minister emphasised that these mechanisms already exist within the current legal structure, suggesting that Malaysia does not anticipate imminent wholesale reform of copyright law specifically targeting AI.
However, Malaysia's regulatory posture appears notably cautious compared to the aggressive expansion of AI development in wealthier nations. The government has not undertaken any dedicated assessment of how artificial intelligence generated content might displace workers or depress earnings within Malaysia's creative sectors, a gap that Senator Wan Martina Wan Yusoff highlighted in supplementary questioning. This omission is significant given the rapid advancement of generative AI tools capable of producing written content, visual media, and music that could compete directly with human creators' output.
Fuziah acknowledged this knowledge gap explicitly, indicating that the government remains in a consultative phase rather than implementation phase regarding AI regulation. Officials are conducting ongoing engagement with stakeholders—a formulation that typically includes industry representatives, civil society groups, and technical experts—while simultaneously monitoring approaches adopted by other nations. This observational strategy reflects a pragmatic recognition that AI technology evolves far more rapidly than legislative processes can accommodate, making premature or overly prescriptive regulation potentially counterproductive.
A notable absence in Malaysia's enforcement landscape is the complete lack of court cases involving copyright infringement stemming from AI training or deployment. This contrasts sharply with jurisdictions like the United States, where major litigation between AI companies and rights holders has already commenced. The absence of Malaysian cases suggests either that AI usage of copyrighted material has not yet reached scales that trigger disputes, or alternatively that creators and rights holders lack sufficient awareness or resources to pursue legal challenges. Neither scenario provides comfort for policymakers concerned about creative industry protection.
Malaysia's copyright framework does contain a critical limitation that shapes AI development opportunities within the country: only works demonstrating originality achieved through human effort, skill, and intellectual contribution receive legal protection. Conversely, material generated entirely by artificial intelligence without substantial human creative input falls outside the copyright system's scope. This distinction creates legal space for AI companies to use their own generated outputs without restraint, though it simultaneously raises questions about whether such machine-generated content should populate training datasets intended for further AI development.
The tension between these positions—requiring permission to use existing copyrighted works while excluding purely machine-generated content from protection—reveals underlying uncertainties about how intellectual property doctrine should evolve alongside AI capabilities. International jurisdictions are grappling with analogous questions about ownership, originality, and protectability in AI contexts. Malaysia's decision to maintain existing law rather than preemptively legislate represents one approach, though it potentially leaves gaps as technology advances beyond current statutory frameworks.
For Malaysia's creative industries, the emphasis on copyright owner consent theoretically offers protection, yet practical enforceability remains questionable. Many Malaysian creators lack the resources or legal sophistication to monitor whether their works are being incorporated into distant AI systems, particularly when those systems operate across international borders. Similarly, many small publishers, musicians, and visual artists may prove inadequately represented through Collective Management Organisations, leaving individual negotiating power fragmented and weak against technologically sophisticated firms backed by substantial capital.
The government's stated intention to continue studying international models before considering new policies suggests that Malaysia will likely move toward revised copyright frameworks within the next few years. Countries including the European Union, United Kingdom, and others have already introduced or proposed specific provisions addressing AI training and copyright, establishing precedents that Malaysian policymakers can evaluate. However, Malaysia's approach appears deliberately unhurried, prioritising understanding before action rather than rushing toward regulation that might later require amendment.
This cautious stance carries both advantages and risks. On one hand, careful deliberation reduces the likelihood of poorly designed rules that hamper legitimate innovation or create unintended consequences. Conversely, the delay permits rapid expansion of AI capabilities and business models that may become difficult or costly to regulate retroactively. Malaysian creators operating in growing digital markets may find themselves disadvantaged during this interim period, with their works absorbed into training datasets while compensation mechanisms remain underdeveloped and enforcement mechanisms untested.
The broader Southeast Asian context adds another dimension to Malaysia's choices. Neighbouring countries are watching how Malaysia addresses AI governance, particularly regarding creative rights protection. Singapore's more permissive approach contrasts with Malaysia's stated preference for copyright owner approval, potentially creating competitive pressures that influence Malaysia's eventual policy direction. Regional harmonisation of AI governance could eventually emerge, but currently each nation is charting its own course based on domestic priorities and international influences.
