Independent course prototype · May be incomplete or outdated · Check linked primary sources
← Back to map
instrument · china

Artificial Intelligence Safety Governance Framework, Version 2.0

A state-backed technical blueprint classifying AI risks and recommending lifecycle measures for technical, application and broader societal risks, including severe loss-of-control scenarios.

currentverified from 2026-07-27non-binding technical framework
Attribute evidence

What the atlas claims

What it is

factualhigh confidencereview A

A state-backed technical blueprint classifying AI risks and recommending lifecycle measures for technical, application and broader societal risks, including severe loss-of-control scenarios.

The document presents risk categories, grading concepts and governance measures.

Legal force

factualhigh confidencereview A

non-binding technical framework

The release describes it as a TC260 technical document rather than a law or administrative regulation.