The Decree governs notification, labeling, incidents, high-risk conformity and impact assessment, the national AI portal, sandboxes, inspection, suspension and withdrawal.
Regulation
Verified record
- Risk classification summary
- AI systems are classified as low, medium or high risk; high-risk systems receive predeployment conformity and impact-assessment duties.
- Requirements summary
- Providers must classify AI systems before use, notify medium- and high-risk systems through the national portal, assess high-risk systems before deployment and after material changes, publish assessment results, label deepfakes, and report serious incidents preliminarily within 72 hours or five business days and fully within 15 days.
- Affected entities
- AI system providers, deployers and organizations responsible for systems or reportable incidents within the Decree’s scope, including public-impact and high-risk systems.
- Enforcement summary
- Competent authorities may require reassessment or inspection and may suspend or withdraw systems in accordance with the Artificial Intelligence Law and the Decree.