AI Basic Act

The Legislative Yuan passed the third reading of the “AI Basic Act” today, designating the NSTC as the central competent authority and local governments as the local competent authorities. The act establishes seven principles for AI development: sustainability and well-being, human autonomy, privacy protection and data governance, information security and safety, transparency and explainability, fairness and non-discrimination, and accountability.

Key Highlights of the Passed Act:

• Governance Structure: The Executive Yuan will establish a “National AI Strategy Special Committee,” convened by the Premier, to set national AI development guidelines.
• Risk Management: MODA will promote a risk classification framework; high-risk AI products or systems must clearly display precautions or warnings.
• Rights Protection: The government shall prevent AI applications from infringing on life, liberty, or property, and actively use AI to ensure labor rights and bridge skill gaps.
• Regulatory Adaptation: The government shall complete the formulation, amendment, or abolition of relevant regulations within 2 years of implementation.

View LawTrace

The parts of Article 18 reserved during consultation by the President reached a consensus among the three parties in the 2025.12.19 plenary session, following the KMT-TPP re-amended motion. The rest passed without objection or textual amendment, completing the third reading. The biggest divergence was on the competent authority; ultimately, the version pushed by the opposition parties passed the vote, designating the National Science and Technology Council (NSTC) as the competent authority, avoiding the predicament of the Executive Yuan’s draft having no competent authority and no national-level strategic committee. The Premier will serve as the convener of the National AI Strategy Council and must promote the development guidelines.

On 2025.12.19, the Legislative Yuan voted to pass Article 2 of the “AI Basic Act” draft regarding the competent authority, which was reserved during the President’s consultation. The KMT and TPP caucuses jointly proposed an amendment and a re-amendment motion, confirming the NSTC as the central competent authority and local governments as local competent authorities.

Legislative Speaker Han Kuo-yu convened a cross-party negotiation on the “AI Basic Act” draft. After discussions, a high degree of consensus was reached, with only the article regarding the “Competent Authority” reserved for a vote in the plenary session; all other articles achieved consensus. During the meeting, Speaker Han affirmed the importance of this bill for Taiwan’s AI development and commended the efforts of the promoting committee members. Latest Negotiated Draft Version

The Executive Yuan’s version of the Artificial Intelligence Basic Act was submitted to the Legislative Yuan and proceeded directly to the second reading. Regarding the versions previously proposed by various legislators, following committee negotiations during my tenure as convener, 5 articles currently remain without consensus. These have entered the negotiation stage convened by the President of the Legislative Yuan. Once negotiations are concluded, the bill will proceed to the third reading. View Legislative Gazette

The Executive Yuan meeting passed Taiwan’s AI Basic Law draft, compiled and drafted by MODA. The law will not establish an AI competent authority, and in the future, various ministries will formulate implementation laws and guidelines. MODA emphasized that the law balances innovative development and human rights protection. The Executive Yuan version of the AI Basic Law will next be submitted to the Legislative Yuan for review. August 28 Executive Yuan Meeting Report, Executive Yuan Draft, MODA Summary

The Legislative Yuan’s Education and Culture Committee today negotiated the AI Basic Law draft. After 4.5 hours of cross-party discussion, significant progress was made, including articles on legislative purpose, government principles for promoting AI R&D and applications, and government’s active use of AI to ensure workers’ labor rights, reaching preliminary consensus. View Legislative Gazette, View Legislative Livestream, Ko Version Amendment Motion - KMT Caucus Integrated Version, Amendment Motion - DPP Caucus Integrated Version, Complete Negotiation Conclusions, WORD | Negotiated Passed Articles

For the first time in Taiwan’s history, under Committee Chair Ko JC at the Legislative Yuan’s Education and Culture Committee joint Transportation Committee, the AI Basic Law draft was scheduled for review. After 2.5 hours of article-by-article discussion, committee review was finally completed on August 4, sending it to the plenary session for processing after negotiations. View Legislative Gazette, View Legislative Livestream, Ko Version - KMT 12 Proposing Legislators Integrated Version - 31 Legislators Proposing Co-signed Amendment Motion

MODA responded to the Legislature with an official document on July 14, but failed to provide the promised ministry draft articles, only offering “opinions” on legislators’ versions. The document contained 14 instances of “requires further consideration” and 3 instances of “suggest waiting for Executive Yuan version before review.” The scheduled review was forced to postpone. View MODA Response Document

Despite MODA having sent the document to the Legislature at 3 PM on July 14, Deputy Minister Lin Yi-jing told media after a press conference on July 15 that it was “still being processed internally and not yet sent out.” The MODA document contained no “ministry draft articles” whatsoever, only opinions on legislators’ versions.

Legislator Ko pressed the Executive Yuan to submit the AI Basic Act on schedule, and announced that an attached resolution would require the Executive Yuan to complete a Child Impact Assessment within a specified timeframe after the Act passes its third reading.

AI development didn’t start in recent years; as early as 2018, when Lai Ching-te was Premier, the “Taiwan AI Action Plan” was proposed. At the same time, Legislator Hsu Yu-jen in the 9th Legislative Yuan also proposed an AI Basic Law draft. Since AI technology has been discussed for years, why hasn’t the government conducted a child and youth impact assessment in the past decade? I believe all 113 legislators, regardless of party, care deeply about children’s rights. However, simply giving administrative agencies more time is irresponsible, condones administrative laziness, and undermines the Legislative Yuan’s legislative authority. As long as the AI Basic Law isn’t passed in third reading, children’s rights remain unprotected.

Therefore, we have resolved to require MODA to submit its ministry version of the articles by July 15, and I will push for an attached resolution mandating a child impact assessment within a certain period after third reading.

National Affairs Forum Video:

Related Resources:

The head of the “AI Top 10 Projects” initiative could not answer basic questions about the bill’s progress, highlighting the administration’s lack of coordination.

From “AI Cabinet,” “Sovereign AI” to “AI New Top 10 Projects,” policy details remain unseen, raising doubts about the government’s true commitment.

The NDC-led “AI New Top 10 Projects” claims to drive NT$15 trillion in output and create 500,000 high-paying jobs, but when questioned on specifics, the Chair said details could only be provided “after Executive Yuan approval.” The plan content, work allocation, and funding are undecided, showing inadequate preparation.

This endless waiting also applies to the AI Basic Law, still pending in the Executive Yuan. Shockingly, the NDC Chair mistakenly thought the draft was already in the Legislature, indicating no grasp of the process. Even more absurdly, NDC, responsible for AI projects, hasn’t participated in bill reviews—how can it lead national AI policy?

NDC Interpellation Video:

Related Resources:

Submission of the All Legislators and TPP Version - AI Basic Law Draft.

Legislator Ko livestreamed to expose the obstruction and called for swift passage.

DPP legislators argued that “AI evolves too quickly,” nitpicked wording without offering amendments, and forced the meeting to adjourn. MODA promised to deliver its Executive Yuan draft to the Legislature by July 15.

Although there is cross-party consensus on “AI’s rapid development,” the DPP argues “due to its speed, we should continue observing international attitudes,” and MODA has given no suggestions on existing versions.

DPP members repeatedly claim the current version lacks protections for women, diversity, disabled persons, indigenous peoples, rural areas, etc., but Articles 4 (“Freedom Protection”), 5 (“Digital Equality”), 6 (“Fraud Prevention”), and more already include general provisions, consistent with basic law principles. Nonetheless, DPP members use this as a pretext for “superficial discussion, actual obstruction.”

DPP members themselves said in the meeting they “hope to let AI develop more” and “MODA should explain why no Executive draft after multiple discussions,” but MODA hasn’t opined on legislator versions or submitted its draft, becoming an excuse for DPP to block review. The MODA Minister cited international observation for delay, ignoring Taiwan’s lack of legislation compared to mature systems in Europe and the US: Countries like the UK avoid over-regulation but are preparing comprehensive norms, not halting legislation.

Internationally, Korea, Japan, UK are actively laying out AI frameworks; if Taiwan continues to dither, it will miss opportunities and expose public rights to risks. Eight years of waiting shouldn’t lead to more delays and empty words.

Committee Review Livestream:

Related Resources:

The Executive Yuan’s Feb 26 decision shifted competent authority from NSTC to MODA with only a 54-character note; NSTC Minister Wu admitted prior ignorance, and MODA has yet to convene any consultations mandated by the draft.

Interpellation Video:

Bill Introduction Video:

Related Resources:

Legislators questioned whether the initial “R&D-focused, regulation as support” approach would shift to a “regulation-oriented” one; raised concerns over the Ministry of Digital Affairs’ capacity (as a third-tier agency) in terms of staff and budget; and criticized the opaque decision-making process, demanding the Executive Yuan clarify its decision-making transparency. Questions also addressed open data, sovereign AI, AI training databases, and energy policy (including nuclear power development).

MODA stated that, as instructed by the Executive Yuan on February 26, the draft is under MODA’s leadership, now placing more emphasis on its regulatory role: “balancing the encouragement of innovation with regulation.”

• CNA Report Supplement: On April 23, MODA further explained that the draft would be responsive to the needs of Taiwan’s industry and society, adhering to the core principles of “encouraging innovation while upholding human rights,” and guiding government agencies to develop and promote AI applications. The draft is not a “regulatory law” per se, but rather sets out principles for AI development and policy integration. Detailed regulations (such as data management and AI risk classification) will be advanced later through subordinate legislation.

Most experts advocated that the competent authority should have horizontal coordination abilities (such as a special committee at the Executive Yuan level), accelerate legislation on sovereign AI and open data, and defend Taiwan’s local language data and cultural sovereignty.

He acknowledged that authority over the AI law had been transferred to MODA, saying, “NSTC will focus on AI research.” The decision was made after inter-ministerial discussions at the Executive Yuan.

Submission of the Ko JC and 37 Legislators Version - AI Basic Law Draft.

The Executive Yuan, without a public resolution at its formal meeting, suddenly instructed that the “draft bill would be overseen by the Ministry of Digital Affairs (MODA).”

• Question: February 26 was not an official Executive Yuan meeting day, raising concerns about decision-making transparency.
• CNA Report Supplement: According to a CNA report on April 23, 2025, MODA explained that the Executive Yuan formally instructed, on February 26, that MODA would be responsible for promoting legislation and providing further interpretation of the AI Basic Law draft. The draft is expected to be submitted to the Legislative Yuan for review after approval by the Executive Yuan. The draft was originally proposed by the National Science and Technology Council (NSTC) last year, made available for public comment from July 15 to September 13, 2024, and reviewed at an Executive Yuan meeting on November 22, 2024, with a request for revisions before resubmission.

The Executive Yuan Deputy Minister explained, “The draft of the AI Basic Law has been submitted for review by the Executive Yuan and will be revised promptly and submitted for legislative review before the next session.”

Legislators questioned whether the draft had been significantly revised by the Executive Yuan. NSTC responded that “the direction has not changed significantly, and at the latest, the draft will be submitted to the Legislative Yuan in the first session next year.”

NSTC submitted the AI Basic Law draft to the Executive Yuan for review.

The public comment period for the draft ended. Minister Wu Cheng-Wen stated in the legislature that regulatory review was complete, and the draft was expected to be sent to the Executive Yuan by the end of October, aiming for legislative review in the same session.

NSTC officially released the draft of the AI Basic Law to the public and opened it for comment.

NSTC convened a consultation meeting on the AI Basic Law draft, inviting industry and human rights groups, and emphasizing a balance between innovation and human rights.

responded in the Legislature: “The AI Basic Law will be launched this year, with an aim to submit it to the Executive Yuan before the end of October.”

Wu Cheng-chung: “We expect to introduce the AI Basic Law by the end of the year.”

Wu Cheng-chung explicitly stated that the draft will follow the models of the US, Japan, and the UK, and will be proposed before year’s end.

National Science and Technology Council (NSTC) Minister Wu Cheng-chung reported in the legislature that an “Executive Yuan version of the AI Basic Law draft” will be available by the end of the year, highlighting that Taiwan will adopt a US/Japan-style “guidance first, legislation later” approach to encourage innovation and monitor international developments.

The Executive Yuan approved the “Taiwan AI Action Plan 2.0 (2023-2026),” which emphasizes AI ethics and legislation. The plan aims to promote AI legal frameworks, establish AI product/system evaluation centers, and develop AI ethics guidelines that interface with international standards, continuing previous initiatives while strengthening Taiwan’s AI industry development and governance framework.

Newly inaugurated Legislator Ko Ju-Chun immediately questioned the progress of the “AI Basic Law.” Then-Premier Chen Chien-jen and Minister without Portfolio Wu Cheng-chung both responded positively, emphasizing “optimism, technological policy continuity, following the US and Japan models, and encouraging R&D.”

Legislative Yuan member Hsu Yu-jen and 21 other legislators drafted the “Artificial Intelligence Development Basic Law,” drawing on the legal and policy experiences of AI-advanced countries including the United States, United Kingdom, and Japan, while integrating Taiwan’s national conditions. The main content includes determining the position of artificial intelligence in national policy, establishing development guidelines, setting up competent authorities and special committees, government infrastructure, data and privacy protection, education budget, sandbox programs, development principles and ethics, aiming to respond to the global Industry 4.0 trend and promote national-level AI development strategy.

The Executive Yuan approved the “Taiwan AI Action Plan” (2018-2021), initiated by then-Premier Lai Ching-te, with a planned annual investment of nearly NT$10 billion. The plan aims to strengthen Taiwan’s existing advantages by combining talent, ICT and semiconductor industries, open fields and data, promoting four major strategies: “AI Talent Sprint,” “AI Leading Promotion,” “Building International AI Innovation Hub,” and “Regulatory and Field Deregulation,” to make Taiwan an AI innovation hub and global smart technology powerhouse, while accelerating the 5+2 industrial innovation.

Then-Premier Lai Ching-te emphasized at the “2018 CommonWealth Economic Forum” that Taiwan faces challenges of declining birth rates and technological development, and the government needs to use policy tools to cultivate technological foundation, mentioning the promotion of AI action plans to address related issues.

Then-Minister of Science and Technology Chen Liang-ji led the establishment of the “AI Action Plan,” serving as a guiding principle for the next five years of AI development in Taiwan’s industry, government, academia, and research sectors, laying the foundation for Taiwan’s AI development.