Evidence
China and NIST moved on agent standards while EU harmonized standards slipped
| Date | Case | Country | What happened | Source |
|---|---|---|---|---|
| Oct 5, 2025 | GB/T 46347-2025 | China | First national AI risk-management maturity standard | Source |
| Nov 2025 | BSI accredited for ISO/IEC 42001 | United Kingdom | First certifier accredited by UKAS | Source |
| Feb 17, 2026 | NIST AI Agent Standards Initiative | United States | Three pillars: industry standards, open protocols, agent security and identity | Source |
| Mar 2026 | IETF CATALIST meeting | International | Exploratory meeting on agent-to-agent and model-to-tool protocols | Source |
| Apr 14, 2026 | China AI standards subcommittee workshop | China | Reviewed 10 drafts | Source |
Rules
Standards are voluntary until a law points to them
| Rule | Jurisdiction | What it requires | Status |
|---|---|---|---|
| ISO/IEC 42001 | International | Certifiable AI management system | Published 2023 |
| EU AI Act harmonized standards (CEN-CENELEC JTC 21) | EU | Presumption of conformity for high-risk AI | Late 2026 at the earliest |
| MIIT AI standardization guideline | China | 50+ national and industry AI standards by 2026 | Issued 2024 |
| GB 45438-2025 AI content labeling | China | Mandatory AI content labels | In force Sep 1, 2025 |
| NIST CAISI AI Agent Standards Initiative | US federal | Agent security, identity and interoperability | Since Feb 2026 |
Status colors: green in force, blue voluntary standard or framework, amber adopted and phasing in, gray draft or pending, red withdrawn.
Record
Page Record
- Datasets
- ai_governance_evidence.csv (5 rows); ai_rules.csv (245 rows)
- Key sources
- GB/T 46347-2025
BSI accredited for ISO/IEC 42001
NIST AI Agent Standards Initiative
IETF CATALIST meeting
China AI standards subcommittee workshop - Last verified
- October 2026