AI Watch: Global regulatory tracker - Taiwan (UPDATED)

[co-authors: Benjamin Li, Eddie Hsiung]

Taiwan's Artificial Intelligence Basic Act provides guiding principles but does not impose obligations on the private sector. Additionally, draft laws are under consideration and sector-specific initiatives are already in place.

Laws/Regulations directly regulating AI (the “AI Regulations”)

On January 14, 2026, the Artificial Intelligence Basic Act (the "AI Basic Act")¹ was officially promulgated and entered into force. The AI Basic Act serves as the legal basis for regulating and promoting AI applications in Taiwan, and sets forth relevant legal standards and guiding principles. However, it does not impose any obligations on the private sector.

Additionally, the Taiwan Government has published multiple guidelines and plans for AI development, such as:

• The AI Technology R&D Guidelines (AI R&D Guidelines) (published in September 2019)²
• The Taiwan Artificial Intelligence Action Plan 2.0 (2023-2026) (AI Action Plan) (approved by the Executive Yuan in April 2023)³

In August 2023, the Executive Yuan also issued guidelines for the use of generative AI by the Executive Yuan and affiliated Government authorities/agencies.⁴

Status of the AI Regulations

In addition to the AI Basic Act, guidelines and plans, such as the AI R&D Guidelines and the AI Action Plan, have already been approved and published (as discussed above).

Other laws affecting AI

There are various laws that do not directly seek to regulate AI but may affect the development or use of AI in Taiwan. A non-exhaustive list of key examples includes:

• The Personal Data Protection Act⁵
• The Financial Technology Development and Innovative Experimentation Act⁶
• The Unmanned Vehicles Technology Innovative and Experimentation Act⁷
• The Medical Devices Act⁸

Intellectual property laws may affect several aspects of AI development and use.

Definition of “AI”

According to the AI Basic Act, AI refers to a system with autonomous operational capabilities. Such a system, through input or sensing, and by means of machine learning and algorithms, can generate outputs—such as predictions, content, recommendations, or decisions—that, for explicit or implicit objectives, affect physical or virtual environments.

Territorial scope

The AI Basic Act does not explicitly specify its territorial scope, but by interpretation, it should apply to Taiwan.

Sectoral scope

The AI Basic Act does not have any sector focus. However, as the AI Basic Act is an indicative and principles-based law, the respective competent authorities for specific sectors will review, amend, or abolish existing relevant laws, regulations and administrative measures, issue relevant interpretations, or further formulate supporting guidelines and implementing regulations.

In addition, legislators in Taiwan have promulgated certain sector-specific laws for companies to test the use of certain technologies (including AI) in controlled regulatory environments, such as:

• The Financial Technology Development and Innovative Experimentation Act (to promote fintech services and companies)⁹
• The Unmanned Vehicles Technology Innovative Experimentation Act (to promote the use of autonomous vehicles)¹⁰

Additionally, in the financial services industry, the Financial Supervisory Commission (Taiwan's financial regulator, the FSC) published the "core principles and policies for use of AI in the financial industry" (FSC AI Policy) in October 2023.¹¹ In June 2024, the FSC also published a set of guidelines for the use of AI in the financial industry.¹²

Compliance roles

The AI Basic Act does not impose any obligations on the private sector. However, in the future, sector-specific competent authorities may impose AI-related compliance obligations.

Furthermore, AI systems are not currently recognized as separate legal entities in Taiwan, so they cannot be subject to civil liability on a standalone basis. An injured party would therefore need to prove that the use of AI technology falls within a category of tort that is recognized under Taiwan's Civil Code and/or the Consumer Protection Act. For example, an injured party may be able to claim that: (i) the manufacturer of a self-driving car liable under Article 7 of the Consumer Protection Act; or (ii) the user of a self-driving car liable under Article 184 of the Civil Code.¹³

Core issues that the AI Regulations seek to address

The AI Basic Act seeks to address several issues. Certain key provisions include:

• The Government shall avoid the application of AI products or systems that infringe upon the life, body, liberty, or property of the people; disrupt social order, national security, or the ecological environment; or involve deviations, discrimination, false advertising, misinformation, fabrication, or other circumstances that violate relevant laws and regulations
• The Government shall abide by the principle of promoting the best interests of children and adolescents. Where AI products or systems are determined to be high-risk applications by the central competent authority for the relevant business, in consultation with the Ministry of Digital Affairs (MODA), clear indications of precautions or warnings shall be provided
• The Government shall actively promote the research and development, application, and infrastructure of AI
• When developing, training, testing, and verifying the impact of emerging AI technologies, the Government shall provide reasonable use, support, and subsidy measures, and shall improve the regulations governing the research, development, and application of AI
• The Government shall establish mechanisms for data openness, sharing and reuse, to enhance the availability of data for AI use, and shall regularly review and adjust relevant laws and regulations
• Each sector-specific competent authority, in consultation with the competent authority for personal data protection, shall avoid unnecessary collection, processing, or use of personal data during the research, development, and application of AI
• The Government shall provide tailored employment counseling to those unemployed as a result of AI utilization

Risk categorization

Under the AI Basic Act, MODA is responsible for developing an AI risk-classification framework that aligns with international systems and for assisting the sector-specific competent authorities in developing risk-based regulatory standards. Such standards should reflect the need for appropriate risk management of AI applications and align with the aforementioned risk-classification framework. The competent authorities shall also assist relevant industries in formulating their own industry guidelines and codes of conduct.

In addition, the Government shall clearly define the attribution of responsibility and the conditions for liability regarding high-risk AI applications, and shall establish mechanisms for remedy, compensation, and insurance.

Key compliance requirements

The AI Basic Act proposes seven guiding principles for AI development:

• Sustainability
• Human autonomy
• Privacy protection and data governance
• Information security and safety
• Transparency and explainability
• Fairness and non-discrimination
• Accountability

Additionally, the AI R&D Guidelines set out the following eight principles that participants in AI technology should be aware of when conducting relevant activities:¹⁴

• Common good and well-being
• Fairness and non-discrimination
• Autonomy and control
• Safety
• Privacy and data governance
• Transparency and traceability
• Explainability
• Accountability and communication

Regulators

The following authorities oversee different aspects of AI:

• For the AI Basic Act, the National Science and Technology Council serves as the central competent authority, while government of the special municipality and county and city governments act as local competent authorities. Sector-specific matters remain within the jurisdiction of their respective regulatory bodies
• The Executive Yuan (Taiwan's Cabinet) will establish a National AI Strategy Special Committee to coordinate national AI policy, supervise implementation, and formulate the National AI Development Program
• The Ministry of Digital Affairs provides guidance and incentives for interdisciplinary digital innovation of AI and other digital economy related industries
• Sector-specific domains involving the use of AI (e.g., healthcare, transportation, finance, etc.) are regulated by relevant bodies such as the Ministry of Health and Welfare, the Ministry of Transportation and Communication, and the Financial Supervisory Commission

Enforcement powers and penalties

The AI Basic Act does not itself impose any obligations on the private sector. Accordingly, enforcement and penalties relating to creation, dissemination and/or use of AI are governed by related violations in non-AI legislation. However, as mentioned above, new compliance obligations related to AI may arise in the future.

1 Please see here (Mandarin only).
2 Please see here (Mandarin only).
3 Please see here.
4 Please see here (Mandarin only).
5 Please see here.
6 Please see here.
7 Please see here.
8 Please see here.
9 Please see here.
10 Please see here.
11 Please see here (Mandarin only).
12 Please see here (Mandarin only).
13 Please see here.
14 Please see here (Mandarin only).

[View source.]