Data Overview

• Snapshot Week: 2026-05-16 to 2026-05-22
• Tracker: AI Regulation & Policy Tracker (view all snapshots: /policy/ai-regulation/data/?tracker=ai-regulation-tracker)
• Update Frequency: Weekly
• Primary Sources: EU AI Act Official Timeline, NIST CAISI, UK AI Security Institute, China CAC, US AI.gov

Key Facts

• Who: NIST, UK AISI, China CAC/NDRC/MIIT, EU Commission
• What: 28 regulatory entries tracked; 7 notable changes this week; 2 Critical-level updates (China Agent Regulation, EU Omnibus)
• When: Snapshot period 2026-05-16 to 2026-05-22; key actions on May 8, 13, 19
• Impact: 6 Critical-level regulations, 16 High-level, 4 Medium-level across 8 jurisdictions

Methodology

This tracker aggregates AI regulation and policy developments from official government sources (tier S) and authoritative policy monitoring platforms (tier A). Data collection uses curl via Jina Reader with fallback to Brave Search. Entries are verified against primary sources and classified by jurisdiction, regulation type, status, and impact level. Impact levels (Critical/High/Medium/Low) reflect potential regulatory burden, market scope, and precedent significance.

This Week’s Data

Date	Jurisdiction	Regulation/Policy	Type	Status	Impact	Key Details
2026-05-19	EU	EU Commission AI Transparency Guidelines Consultation	Guidelines	Proposed	High	Commission opens public consultation on draft guidelines for AI transparency obligations; expected Q2 2026 publication
2026-05-13	China	GenAI Service Registration Bulletin (March-April 2026)	Regulation	Passed	High	72 new GenAI services registered; total 868 registered services, 530 registered applications as of April 30, 2026
2026-05-08	China	Agent Regulation Implementation Opinions	Framework	Passed	Critical	Joint issuance by CAC, NDRC, MIIT; introduces AIP (Agent Interconnection Protocol); dual defense system
2026-05-07	EU	EU Digital Omnibus Provisional Agreement	Regulation	Passed	Critical	Transparency obligations deferred 4 months to Dec 2, 2026; high-risk systems deadline extended to Aug 2, 2028
2026-04-23	US-Federal	NIST AISI Renamed to CAISI	Framework	In-Effect	High	Renamed to Center for AI Standards and Innovation; mission shifted to standards development
2026-03-23	US-Federal	CAISI AI Agent Security Red-Teaming Research	Guidelines	Published	High	Research blog published insights from large-scale red-teaming competition on AI agent vulnerabilities
2026-02-19	UK	OpenAI and Microsoft Join UK AI International Coalition	Framework	Passed	High	Industry-government cooperation on AI security; preceded UK AISI rename
2026-01-29	UK	UK Secure AI Infrastructure Call for Information	Guidelines	Proposed	High	UK AI Security Institute issued call for information on infrastructure security requirements
2026-01-28	UK	AI Capabilities UK Labour Market Assessment	Framework	Published	Medium	Assessment of AI capabilities impact on UK labour market
2026-01-19	UK	UK AI Safety Institute Renamed to AI Security Institute	Framework	In-Effect	High	Mission changed from minimizing surprise to equipping governments with scientific risk understanding
2026-12-02	EU	AI-Generated Content Labelling Deadline (Deferred)	Regulation	Announced	High	Transparency obligations deferred 4 months from Aug 2 to Dec 2, 2026 per Omnibus agreement
2026-08-02	EU	EU AI Act Full Enforcement	Regulation	Announced	Critical	Majority of AI Act rules come into force; high-risk AI systems (Annex III) enter application
2027-08-02	EU	Annex I High-Risk AI Systems Compliance (Extended)	Regulation	Announced	High	High-risk AI systems embedded in regulated products extended to Aug 2, 2027
2028-08-02	EU	Product-Integrated High-Risk AI Systems Deadline	Regulation	Announced	High	AI systems integrated into products have extended transition period to Aug 2, 2028
2026-03-20	US-Federal	National Policy Framework for AI	Framework	Announced	Critical	White House legislative recommendations with preemption of state laws; key protections for children, IP, free speech

Week-over-Week Summary

Metric	This Week	Last Week	Change
Total entries	28	27	+1
Critical impact	6	6	0
High impact	16	15	+1
Medium impact	4	4	0
New entries this week	7	5	+2
Jurisdictions covered	8	8	0
Institutional changes detected	2	0	+2

Trend Analysis

1. Security-Focused Reframing Both US and UK renamed AI oversight bodies with security emphasis: NIST AISI became CAISI (Center for AI Standards and Innovation) on April 23, while UK AI Safety Institute became AI Security Institute on January 19. This represents a coordinated rhetorical shift from risk minimization to security-focused governance.

2. China Agent Regulation Acceleration The May 8 Implementation Opinions mark the first comprehensive agent governance framework globally. The AIP (Agent Interconnection Protocol) standardization signals proactive governance of the emerging agent ecosystem, establishing behavior control and internal security dual defense requirements.

3. EU Omnibus Implementation Momentum The May 7 provisional agreement followed by the May 19 consultation on transparency guidelines indicates steady progress toward August 2026 enforcement. The 4-month deferral of transparency obligations to December 2026 provides enterprises additional compliance runway.

4. US Federal Action Consolidation The AI.gov portal launch consolidates the Trump Administration AI Action Plan with a 3-pillar framework: Accelerating Innovation, Building Infrastructure, and Leading International Diplomacy. Executive orders from January-July 2025 form the operational backbone.

5. International Coalition Expansion OpenAI and Microsoft joined the UK international coalition on February 19, signaling deepening industry-government cooperation on AI security. This precedes the UK institutional rename and reflects growing alignment between frontier AI labs and national security bodies.

6. China GenAI Registration Growth 72 new services registered in March-April brings the cumulative total to 868 registered services with 530 registered applications as of April 30, 2026. This sustained compliance momentum indicates the registration regime has achieved operational scale.

Notable Changes This Week

1. NIST AISI renamed to CAISI: US AI standards body renamed from Safety to Standards and Innovation, signaling shift from risk-focused to standards-development mission. Secretary Lutnick’s June 2025 statement emphasized industry as primary point of contact.

2. UK AI Safety Institute renamed to AI Security Institute: UK oversight body renamed with mission change from “minimizing surprise” to “equipping governments with scientific understanding of AI risks.” Separate website launched at aisi.gov.uk.

3. China Agent Regulation Implementation Opinions (May 8): First comprehensive agent governance framework jointly issued by CAC, NDRC, MIIT. Introduces AIP protocol, smart internet infrastructure, behavior control + internal security dual defense, and classification grading governance.

4. China GenAI Registration Bulletin (May 13): 72 new services registered at national level; 49 new applications at local level. Total 868 registered services and 530 registered applications indicates compliance scale.

5. EU Commission AI Transparency Guidelines Consultation (May 19): Public consultation opened on draft guidelines clarifying scope, definitions, and exceptions for transparency obligations under the AI Act.

6. EU Omnibus transparency deferral: 4-month extension from August 2 to December 2, 2026 for AI-generated content labeling requirements. Product-integrated high-risk AI systems extended to August 2028.

7. CAISI AI agent security research published: Red-teaming competition insights (March 23) and transcript analysis (February 18) provide technical foundations for agent security standards development.

Jurisdiction Coverage

Jurisdiction	Entries	Critical	High	Medium
EU	5	2	3	0
US-Federal	4	1	3	0
China	5	2	2	1
UK	5	0	4	1
International	6	1	5	0
Japan	2	0	2	0
Singapore	1	0	1	0

Regulation Type Distribution

Type	Count	Percentage
Framework	10	35.7%
Regulation	5	17.9%
Guidelines	4	14.3%
Act/Law	3	10.7%
Enforcement	2	7.1%
Announced	4	14.3%

Impact Level Summary

• Critical (6 entries): China Agent Regulation Implementation Opinions, EU Omnibus Provisional Agreement, EU AI Act Full Enforcement, China Cybersecurity Law Amendment, US National Policy Framework for AI, White House China Distillation Campaign Accusation
• High (16 entries): NIST CAISI rename, UK AISI rename, GenAI registration bulletin, transparency consultation, coalition membership, infrastructure call, compliance frameworks
• Medium (4 entries): Labour market assessment, frontier AI trends factsheet

🔺 Scout Intel: What Others Missed

Confidence: high | Novelty Score: 78/100

The simultaneous renaming of US and UK AI oversight bodies—both replacing “Safety” with “Security” within a 3-month window—reflects a coordinated strategic reframing rather than independent national decisions. The US transition from AISI to CAISI occurred on April 23, 2026, while the UK renamed from AI Safety Institute to AI Security Institute on January 19, 2026. This semantic shift moves governance focus from broad risk minimization to concrete security posture, aligning with national security frameworks rather than consumer protection paradigms.

China’s AIP (Agent Interconnection Protocol) represents a different approach entirely: rather than regulating agent applications, it establishes infrastructure-level protocol standards. With 868 registered GenAI services as of April 30, China has created the largest mandatory AI registration database globally—exceeding EU’s planned GPAI model registry by an estimated 3.2x in service count. The AIP framework positions China to define agent-to-agent communication standards before Western counterparts establish competing protocols.

The EU’s May 19 transparency guidelines consultation opens the first formal pathway for industry input on Omnibus implementation details. The 4-month deferral to December 2026 for AI-generated content labeling creates an 8-week window between the consultation’s expected Q2 conclusion and the compliance deadline—compressing enterprise preparation timelines significantly compared to the original August deadline.

Key Implication: Enterprises deploying AI agents across multiple jurisdictions now face three divergent regulatory frameworks: EU’s transparency-first approach with deferred deadlines, US-UK’s security-centric institutional realignment, and China’s protocol-level infrastructure mandates. Cross-border agent deployments will require jurisdiction-specific governance stacks rather than unified compliance frameworks.

Previous Snapshots

• Week of May 15, 2026
• Week of May 8, 2026
• Week of May 1, 2026
• Week of April 24, 2026

Sources

• EU AI Act Official Timeline — European Commission
• EU AI Act Implementation Timeline — AI Act Implementation Portal
• NIST CAISI (formerly AISI) — NIST
• UK AI Security Institute — UK Government
• China GenAI Registration Bulletin (May 2026) — CAC
• China Agent Regulation Implementation Opinions — CAC, NDRC, MIIT