Home >> News & Publications >> Newsletter

Newsletter

搜尋

  • 年度搜尋:
  • 專業領域:
  • 時間區間:
    ~
  • 關鍵字:

Draft of AI Basic Act Released: Regulating Development and Application



On August 28, 2025, the Taiwan Executive Yuan announced a draft of the Artificial Intelligence Basic Act ("AI Basic Act"). The draft consists of a total of 17 articles, with Article 1 stating the legislative purpose and Article 17 stipulating that the implementation date shall be determined by the Executive Yuan. The summaries of the remaining articles are as follows:
 
The draft defines artificial intelligence as a self-operating system that applies machine learning and algorithms to generate outputs such as predictions, content, suggestions, or decisions (Article 2), and establishes seven guiding principles for AI development (Article 3):
(1) Sustainability: AI development should balance social equity and environmental sustainability while reducing the digital divide.
(2) Human Autonomy: AI should respect human autonomy, personal rights, and fundamental freedoms, while allowing for human supervision.
(3) Privacy Protection and Data Governance: Personal data privacy must be protected, with a focus on data minimization. The AI Basic Act also promotes the open use and reuse of non-sensitive data.
(4) Safety: Security measures must be in place to ensure the robustness and safety of AI systems against threats and attacks.
(5) Transparency and Explainability: The output of AI should be appropriately disclosed or labeled to help evaluate potential risks and understand the impact on related rights.
(6) Fairness: Algorithms should be developed to avoid bias and discriminatory outcomes against specific groups.
(7) Accountability: Developers and users of AI  should bear  responsibilities aligned with their roles.
 
Regarding policy promotion, the draft requires the government to actively promote AI research, development, and infrastructure (Article 4). It also mandates that government agencies shall establish "innovation sandboxes" to foster technological development (Article 5), and requires the government to engage in public-private partnerships and international collaboration to facilitate the exchange of talent and technology (Article 6). Furthermore, AI learning is to be promoted across the entire education system to enhance public literacy (Article 7).
 
Regarding risk management and social impact, the draft requires the government to prevent AI applications from harming people's lives, property, or national security (Article 8). The Ministry of Digital Development is tasked with creating an AI risk classification framework based on international standards. Relevant government agencies shall then use this framework to create tiered management regulations and restrict or prohibit high-risk applications (Article 9).
 
The draft also stipulates that the government may establish standards, verification, and accountability mechanisms to enhance the verifiability and human controllability of AI decisions (Article 10). For high-risk AI applications, the act requires clarifying liability attribution and establishing remedy, compensation, or insurance mechanisms (Article 11). Concerning labor rights, the government must protect workers' occupational safety and provide reemployment assistance to those who become unemployed due to AI (Article 12).
 
Furthermore, the draft regulates data governance and personal privacy. The government authority in charge of personal data protection should assist other government agencies in avoiding unnecessary collection, processing, or use of personal data during AI development and application, integrating privacy protection into the design of AI systems (Article 13). The government is also required to establish a mechanism for data openness and sharing to increase the availability of data for AI, while upholding diverse cultural values and intellectual property rights (Article 14).
 
Finally, the draft requires government agencies to conduct risk assessments and establish internal controls when using AI for their operations (Article 15). To achieve the AI Basic Act's objectives, agencies must review and adjust their duties and regulations accordingly (Article 16). 
 
This draft bill will be submitted to the Legislative Yuan for review.
 
回上一頁