(CNA, Taipei, Dec. 6) — The Executive Yuan has approved the draft Artificial Intelligence (AI) Basic Law and submitted it to the Legislative Yuan for deliberation. Scholars point out that the draft does not specify a single competent authority. Given that AI development involves cross-ministry coordination, they suggest Taiwan look to the models of Japan and South Korea and establish a dedicated committee or office under the Executive Yuan to integrate cross-domain policy frameworks. Additionally, they emphasize that individual laws, such as the Copyright Act, must be updated quickly to implement the spirit of the Basic Law.

The Executive Yuan passed the draft AI Basic Law in late August. The draft stipulates basic principles for AI research, development, and application, and mandates the avoidance of illegal activities or unnecessary collection of personal data through AI applications. Positioned as a fundamental law, it does not designate a single competent authority; instead, the Ministry of Digital Affairs (MODA) will lead the promotion of an AI risk classification framework. Competent authorities for specific industries will then formulate tiered management regulations based on this framework according to their risk management needs.

Chang Li-ching, Chair Professor at Shih Chien University’s College of Law and CEO of the International Artificial Intelligence and Law Research Foundation, stated that once the Basic Law is enacted, the most crucial aspect is its implementation and substantive effectiveness. The key lies in whether the administrative system possesses a structural framework for cross-domain policy integration. She warned that without a designated authority to coordinate efforts, a "multi-headed carriage" scenario could arise, leading to fragmented and disjointed AI policies.

She pointed out that Japan established an AI Strategy Headquarters within the Cabinet Office, led by the Prime Minister, to formulate basic AI plans and coordinate key R&D and application policies. Similarly, South Korea established a National AI Strategy Committee to effectively deliberate and coordinate major AI policies, ensuring consistency and speed across ministries.

Chang believes that Taiwan currently lacks a high-level coordination mechanism comparable to those in Japan and South Korea. She suggests the Executive Yuan establish a "National Artificial Intelligence Committee" to shoulder the responsibility of cross-ministry coordination. This would allow the policy goals of the AI Basic Law to be truly realized in administrative practice, shaping a more unified and forward-looking national AI policy framework.

Wang Chi-hsuan, Associate Professor in the Department of Law at Chinese Culture University, also suggested referencing the model of the Executive Yuan’s Office of Food Safety. He proposed establishing an "Executive Yuan AI Development Office" to coordinate inter-ministerial AI affairs and inviting experts from industry, academia, and research institutes to participate in Executive Yuan meetings to accelerate the integration of diverse opinions.

Wang noted that the draft AI Basic Law attempts to balance innovation and regulation—for example, by using subsidies and incentives to attract industry R&D while requiring the avoidance of harm to citizens' rights. However, the draft does not specify a deadline for various ministries to review and update relevant regulations or guidelines. Given the rapid pace of AI technological development, he suggests setting clear timelines to ensure regulations can respond to changes in AI applications in a timely manner.

Wang added that the AI Basic Law is merely the beginning of subsequent AI-related legislation and amendments. Laws involving AI must be comprehensively updated to fully address the various issues brought about by the technology. Using the definition of "high-risk AI" in the draft as an example, he noted that definitions will vary depending on the usage context—such as the differences between smart healthcare and autonomous driving. These involve professional judgments from different competent authorities, and discussions should be accelerated.

Chang Li-ching also pointed out that while the draft AI Basic Law provides the policy direction, it is only the start of building a legal framework. Different AI applications face unique challenges and risks that still rely on individual departmental laws for resolution. She cited the Copyright Act as an example: the last major amendment was in early 2022, but ChatGPT emerged later that year, meaning the current law has not yet addressed the impact of generative AI. As disputes arising from AI increase, the Copyright Act must be amended accordingly.

Back to the top