In 2026, the most significant milestone in AI law in Taiwan should be the enactment of the Artificial Intelligence Basic Act (AI Basic Act), which came into effect in January 2026.
Although this statute primarily sets the fundamental principles guiding Taiwan’s AI policy and regulatory framework, it signals the government’s strong commitment to fostering AI development and lays the groundwork for future AI legislation and regulation.
Principles guiding Taiwan’s AI governance framework
The AI Basic Act establishes seven fundamental principles that aim to guide Taiwan’s AI development: sustainable development and well-being; human autonomy; privacy protection and data governance; cybersecurity and safety; transparency and explainability; fairness and non-discrimination; and accountability.
Partner
Lee and Li
Taipei
Tel: +886 2 2763 8000 ext. 2179
Email: kenying@leeandli.com
While the order of these principles does not necessarily indicate their relative importance, it is notable that sustainable development is listed first, followed by human autonomy and privacy protection.
Similar to other jurisdictions, Taiwan emphasises transparency, explainability, fairness and non-discrimination.
It is also worth highlighting that the inclusion of “accountability” indicates a significant development in Taiwan’s legal system.
In Taiwan, this term is frequently used in discussing and establishing a governance framework. However, it is rarely expressly codified in Taiwan law.
Taiwan’s policy framework for AI growth and investment
Legal interpretation. The AI Basic Act mandates the government to establish comprehensive regulations governing the research, development and application of AI.
In cases where conflicts arise between these new AI laws and existing legislation, the promotion of any new technologies and services will always take precedence, provided that the seven fundamental principles are upheld.
Support and funding. The AI Basic Act obliges the government to provide reasonable use measures, assistance and support during the AI develop- ment, training, testing and impact assessment phases.
This suggests that Taiwan may facilitate AI developers’ access to training data through mechanisms deemed “reasonable” or based on the “fair use” provisions under the Copyright Act.
The government is tasked with promoting AI R&D, application and infrastructure, conducting resource planning, and offering support, guidance and incentives to the AI industry, including tax benefits and financial incentives.
To encourage innovation and sustainable growth, sector-specific authorities may establish or enhance experimental environments, such as regulatory sandboxes, to foster AI product and service development, similar to the EU’s approach under the AI Act.
Within its fiscal capacity, the government must allocate sufficient budgetary resources and adopt necessary measures to ensure continuous funding for AI policy implementation and development.
AI risk classification under Taiwan’s AI Basic Act
The AI Basic Act divides AI risks into two categories: high-risk and non-high-risk. High-risk AI applications must carry appropriate warnings and alerts.
While the act hints that certain AI products, services or uses may eventually be deemed “prohibitive”, it does not explicitly specify what these will be.
The Ministry of Digital Affairs (MODA) is empowered to classify AI risks, referencing international standards and norms to develop a risk classification framework.
The MODA will assist sectoral regulators in formulating risk-based AI management policies. Based on this framework, regulators may guide industry self-regulation or establish codes of conduct.
The MODA is also responsible for developing tools and methods to help other regulators assess AI risks.
If a regulator finds that an AI application infringes on life, bodily integrity, liberty or property; disrupts social order, security or the ecological environment; or violates laws related to bias, discrimination, false advertising, misinformation or fabrication, it may restrict or prohibit such applications.
This provision opens the door for sectoral regulators to propose new laws restricting or banning certain AI uses.
Regarding high-risk AI applications, the government must clearly define the elements of liability attribution and establish mechanisms for remedies, compensation or insurance.
Taiwan’s AI data governance framework
The AI Basic Act calls on the government to establish mechanisms for open data, data sharing and data reuse to improve the quality of AI training data in Taiwan.
At the same time, it stresses privacy protection by requiring sectoral authorities to collaborate with privacy regulators to ensure data minimisation throughout AI development.
IP rights and training data in Taiwan
Intellectual property rights can pose challenges for AI training.
Taiwan’s IP authorities have taken a conservative stance on copyright issues related to AI.
In June 2025, a landmark court ruling held that using web-crawling technology to develop database-related services resulted in criminal liability and significant civil damages.
This decision raises concerns about the legality of acquiring training data from the internet.
Beyond imposing an obligation on the government to provide “reasonable use” measures during AI development, the AI Basic Act additionally requires efforts to improve the quality and quantity of the data available for AI as part of its aim to ensure that all training data and outputs reflect Taiwan’s diverse cultural values while also safeguarding IP rights.
AI talent and labour rights protection in Taiwan
The act mandates the continuous promotion of AI and AI ethics education across schools, industries, organisations, society and public institutions to enhance digital literacy. The government is also tasked with advancing AI development policies, and encouraging collaboration among public, private and academic sectors.
This includes fostering interdisciplinary co-operation, facilitating talent and technology exchanges, and supporting the relevant developments in infrastructure.
The government commits to using AI to protect labour rights and interests. It will proactively address the skills gap caused by AI advancements, increase workforce participation, safeguard economic security and uphold labour dignity.
Employment counselling and support will be provided to individuals displaced by AI technologies, tailored to their abilities and needs.
Taiwan’s two-year AI law reform timeline
Under the AI Basic Act, the government must review existing laws, regulations and administrative measures within its authority.
If any are inconsistent with the act or if gaps exist, the government will enact, amend or repeal laws and improve administrative measures within two years from January 2026.
Before enacting or amending laws, if existing legislation lacks relevant provisions, the central competent authority of each sector will interpret and apply the AI Basic Act accordingly.
Outlook
With the AI Basic Act now in force, the Taiwan government has clearly demonstrated its strong commitment to advancing AI development while upholding key principles.
It is expected that additional policies and incentives will be introduced to support AI growth and also benefit Taiwanese companies.
As the MODA develops AI risk classification systems and assists other agencies in assessing AI risks and shaping future policies, new laws and regulations are likely to follow.
The next two years will be a critical period for AI development in Taiwan, both legally and technologically.
Lee and Li, Attorneys-at-Law8F, No. 555, Sec 4, Zhongxiao E Rd,
Taipei 110055, Taiwan
Tel: +886 2 2763 8000
Email: attorneys@leeandli.com
