TAIPEI (Taiwan News) — The Legislative Yuan on Tuesday passed the Artificial Intelligence Basic Act in its third reading, establishing a legal framework for AI governance and naming the National Science and Technology Council as the central competent authority.
Under the law, artificial intelligence is defined as a system capable of autonomous operation that, through inputs or sensing and using machine learning and algorithms, can generate outputs affecting physical or virtual environments, per CNA. These outputs include predictions, content, recommendations or decisions for explicit or implicit objectives.
The act stipulates that government promotion of AI research and applications must balance social welfare, digital equity, innovation, and national competitiveness. It must also adhere to seven core principles: sustainability and well-being, human autonomy, privacy and data governance, cybersecurity and safety, transparency and explainability, fairness and non-discrimination, and accountability.
To prevent harm from AI applications, the law requires the government to avoid uses that infringe on life, bodily integrity, freedom, or property. AI must also not disrupt social order, national security or the environment, or involve bias, discrimination, false advertising, misinformation, or fabrication.
The government must prioritize the best interests of children and adolescents and clearly label high-risk AI products or systems with precautions or warnings. It should also provide or recommend evaluation tools or methods developed in consultation with industry, academia, civil society groups, and legal experts.
The law mandates that the Cabinet establish a national AI strategy committee convened by the premier and comprising scholars, experts, industry representatives, agency heads and local government leaders. The committee must meet at least once a year, with administrative support provided by the NSTC.
The government should allocate sufficient budgets within its fiscal capacity, adopt necessary measures, and actively promote AI research, applications, and infrastructure. It must also provide support, subsidies and assistance when developing, training and testing emerging AI technologies and assessing their impact.
The law further requires strengthening the legal and regulatory framework governing AI research and applications. It also mandates mechanisms for data openness, sharing and reuse to enhance availability for AI use.
Personal data collection, processing or use must be minimized in AI development and application, and data protection by design and by default must be promoted to safeguard individual rights.
The Ministry of Digital Affairs must reference international standards or norms to develop an AI risk classification framework aligned with global practices. It shall establish risk-based management norms and assist industries in formulating sector-specific guidelines and codes of conduct.
The law requires clear liability attribution and conditions for high-risk AI applications, and the establishment of relief, compensation, or insurance mechanisms. Government use of AI in official duties or public services must include risk assessments and appropriate mitigation measures.
The government must actively leverage AI to safeguard workers’ rights, reduce skills gaps caused by AI, and increase labor force participation. It must also provide employment guidance and assistance to workers who become unemployed due to AI adoption, and ensure economic security while promoting decent work.
