AI regulation in Taiwan: laws & policy (2026)
AI in Taiwan: comprehensive law.
Taiwan enacted the Artificial Intelligence Basic Act on 23 December 2025 (promulgated and in force 14 January 2026), establishing a 20-clause framework law that sets seven core principles for AI governance, designates the NSTC as the lead authority, and creates a three-tiered national governance structure. The act is deliberately a framework statute: it articulates principles, institutional roles, and high-risk AI obligations, but delegates penalty and enforcement details to implementing regulations and sectoral legislation to be enacted within two years. It complements Taiwan's AI Action Plan 2.0 and the Executive Yuan's 10 AI Infrastructure Initiatives launched in 2025.
Legal framework
Artificial Intelligence Basic Act (人工智慧基本法), promulgated 14 January 2026; National Science and Technology Council (NSTC) as central competent authority; AI Strategy Special Committee under the Executive Yuan
Key points: artificial intelligence in Taiwan
Enactment & promulgation
The Legislative Yuan passed the AI Basic Act on 23 December 2025; President William Lai promulgated it on 14 January 2026, making it binding law. The act supersedes inconsistent legislation, with a two-year window for government to amend or repeal conflicting laws.
Governance structure
A three-tiered structure places an AI Strategy Special Committee (chaired by the Premier, including academics, industry, agency heads and local government) at the top, with the NSTC as the central executing authority and sectoral ministries below. The committee must convene at least annually.
Seven core principles
The act mandates that all AI research, development and application adhere to: sustainability and well-being; human autonomy; privacy and data governance; cybersecurity and safety; transparency and explainability; fairness and non-discrimination; and accountability.
High-risk AI obligations
The act requires the government to classify AI applications by risk level, mandate warnings/precautions for high-risk systems, and establish liability attribution standards, compensation mechanisms, and insurance frameworks for harms caused by high-risk AI. MODA targeted Q1 2026 to publish the implementing risk-classification framework.
No penalties in base law
The AI Basic Act does not itself include penalties, fines, or detailed compliance procedures; these are deferred to NSTC-led implementing regulations and cross-ministry sectoral rules, which must be in place by approximately January 2028.
National AI strategy context
The act builds on the Taiwan AI Action Plan 2.0 (launched 2023) and the Executive Yuan's 10 AI Infrastructure Initiatives (2025) targeting silicon photonics, quantum computing, and AI robotics, positioning the law within a broader industrial-strategy framework rather than purely a regulatory one.
Timeline of decisions
- DecisionNational AI Strategy Special Committee Formation Announced
Taiwan's Executive Yuan announced it will formally establish the National AI Strategy Special Committee, to be chaired by the Premier and tasked with setting the national AI development blueprint and coordinating AI governance across ministries. This is a mandatory implementation step under the January 2026 AI Basic Act.
Digitimes
- LawPresident Promulgates AI Basic Act, Law Enters into Force
President William Lai Ching-te signed the Artificial Intelligence Basic Act into law, making Taiwan one of the first Asian jurisdictions with a dedicated AI governance statute. The Act designates the NSTC as the central competent authority and imposes a 24-month deadline for all ministries to align subordinate regulations.
Library of Congress — Global Legal Monitor
- Law · officialLegislative Yuan Passes Artificial Intelligence Basic Act (Third Reading)
Taiwan's parliament completed the third reading of the AI Basic Act, enshrining seven core governance principles, sustainability, human autonomy, privacy, cybersecurity, transparency, fairness, and accountability, and designating NSTC as the national AI supervisory authority. The Ministry of Digital Affairs (MODA) was separately tasked with developing an internationally aligned risk-classification framework.
Ministry of Digital Affairs (MODA)
- Decision · officialExecutive Yuan Approves Draft AI Basic Act and Transmits to Legislature
After extensive inter-agency review, the Cabinet cleared the draft Artificial Intelligence Basic Act and forwarded it to the Legislative Yuan, adopting a principles-based framework approach rather than sector-specific prohibitions. The approved draft consolidated competing DPP and KMT legislative proposals that had circulated since late 2024.
Executive Yuan, R.O.C. (Taiwan)
- Guidance · officialNSTC Publishes Draft AI Basic Act for 60-Day Public Consultation
The National Science and Technology Council released its draft Artificial Intelligence Basic Act for public comment, kicking off an 18-month legislative process. The draft aligned with OECD AI Principles and proposed a risk-tiered regulatory model broadly analogous to the EU AI Act.
NSTC — Office of Science and Technology Policy
- DecisionAI Product and System Evaluation Center Established
Taiwan established a dedicated AI Product and System Evaluation Center to develop testing methodologies and certify AI products against ethical and safety standards. The center provides the institutional infrastructure for future risk-based compliance envisioned in AI Action Plan 2.0.
Science & Technology Law Institute (STLI / III)
- Guidance · officialExecutive Yuan Issues Reference Guidelines for Government Use of Generative AI
The Cabinet issued binding reference guidelines requiring government personnel to maintain human oversight when using generative AI tools, prohibiting cloud-based AI (e.g., ChatGPT) for classified or sensitive documents, and mandating accountable human review of all AI-assisted outputs. This was Taiwan's first formal generative-AI governance instrument.
Executive Yuan, R.O.C. (Taiwan)
- Decision · officialAI Action Plan 2.0 (2023-2026) Approved and TAIDE Sovereign-AI Project Launched
The Executive Yuan approved the second multi-year national AI strategy, for the first time elevating AI ethics and legislation as explicit pillars and mandating the NSTC to draft a dedicated AI Basic Act. Simultaneously, the government launched the Trustworthy AI Dialogue Engine (TAIDE), a domestically developed large language model built on Llama incorporating Taiwanese language and values, to reduce dependence on foreign AI infrastructure.
Executive Yuan, R.O.C. (Taiwan)
- Decision · officialTaiwan AI Action Plan 1.0 (2018-2021) Launched, AI Declared a National Priority
The Executive Yuan rolled out Taiwan's first whole-of-government AI strategy, allocating NT$9-10 billion annually to build AI talent pipelines, university research centers, and industrial AI applications, with goals including training 33,000 professionals and opening a self-driving vehicle test site. This plan established AI as a strategic national priority and created the institutional and semiconductor-industry foundations underlying all subsequent AI policy.
Executive Yuan, R.O.C. (Taiwan)
Sources
- focustaiwan.tw/politics/202512230010
- taipeitimes.com/News/front/archives/2025/12/24/2003849407
- loc.gov/item/global-legal-monitor/2026-05-07/taiwan-new-law-establishes-basic
- bakermckenzie.com/en/insight/publications/2026/01/taiwan-ai-basic-act
- iapp.org/news/a/taiwan-s-strategic-leap-into-ai-enacting-the-ai-basic-act-to-f
- whitecase.com/insight-our-thinking/ai-watch-global-regulatory-tracker-taiwan
- english.ey.gov.tw/News3/9E5540D592A5FECD/dfa7107b-f323-4961-a31b-873bbc12ea47
- english.ey.gov.tw/News3/9E5540D592A5FECD/1dec0902-e02a-49c6-870d-e77208481667
- ostp.nstc.gov.tw/PolicyContent.aspx?id=12
- law.asia/taiwan-ai-basic-act-legal-framework-risk-policy/
Next for Taiwan: Data & PrivacyComprehensive law
Next topicOrientation, not legal advice: verify against the primary sources linked above. Last verified 24 May 2026. Methodology & how to cite · State of Technology Regulation 2026