To ensure the safety and reliability of technological applications while encouraging the development of artificial intelligence, the Ministry of Digital Affairs (moda), in accordance with the provisions of the AI Basic Act, is promoting an "AI Risk Taxonomy and Assessment Framework" aligned with international standards. This framework is intended to serve as a common language for inter-ministerial communication within the government and to provide a unified set of assessment standards.

I. Framework Positioning and Objectives

• Common Assessment Standards: Assisting various competent authorities in identifying potential risks of AI applications within their respective jurisdictions and reviewing the adequacy of existing regulations.
• Inter-ministerial Coordination: Assisting competent authorities of various industries in formulating risk-based management regulations.
• Division of Powers and Responsibilities: The ultimate power and responsibility for governance and implementation remain with the respective competent authorities; following the principles of this framework, they will translate these into legally binding and directly enforceable domain-specific management regulations.

II. Four Core Operational Processes

This framework guides various agencies to perform the following operations in sequence to implement risk management:

• AI Use Case Inventory: Gaining an understanding of current or potential future AI application scenarios within the industries under their jurisdiction.
• Risk Identification: Identifying potential risk types within industrial AI applications.
• Risk Assessment: Further evaluating the impact level of the identified risk types.
• Risk Response: Strengthening or planning appropriate management actions based on the determination results.

III. Important Schedule Forecast

The draft of this framework was submitted to the Executive Yuan for review in March 2026.