Regulatory Frameworks

How the EU, China, the US and others regulate AI.

The EU pushed its high-risk AI rules back 16 months to December 2027, while China kept adding sector rules, most recently its May 2026 opinions on AI agents. Korea's AI Basic Act took effect in January 2026 with a top fine of about $21,000; the EU's is €35 million or 7% of turnover.

Maximum fine per violation: EU AI Act, prohibited practices €35M or 7%; China Cybersecurity Law, infrastructure RMB 10M; California SB 53 $1M; Texas TRAIGA $200,000; Korea AI Basic Act KRW 30M. Japan's AI Promotion Act has no penalties MAXIMUM FINE PER VIOLATION EU AI Act, prohibited practices €35M or 7% China Cybersecurity Law, infrastructure RMB 10M California SB 53 $1M Texas TRAIGA $200,000 Korea AI Basic Act KRW 30M US DOLLARS EQUIVALENT, LOG SCALE JAPAN'S AI PROMOTION ACT HAS NO PENALTIES
EU Annex III deadline
Dec 2, 2027
China agent priority scenarios
19
SB 53 compute threshold
10^26 FLOPs
Registry incidents
4

Jurisdictions

China accounts for 62 of the 245 rules 137AI tracks

Rules tracked, by jurisdiction and status
JurisdictionIn forcePhasing inVoluntaryDraft or pendingWithdrawnTotal
China47294062
EU233042160
US federal291118150
US states15600021
UK111205
International53306044
Company policy003003

The full list is in the rules tracker.

Evidence

Every major jurisdiction changed course in 2026

Developments: regulatory Frameworks
DateCaseCountryWhat happenedSource
May 8, 2026China agent Implementation OpinionsChinaThree decision tiers plus filing, testing and recall for agents in sensitive sectorsSource
Jan 22, 2026Korea AI Basic Act takes effectSouth KoreaHigh-impact AI duties, generative AI labeling, a local representative for large foreign firmsSource
May 14, 2026Colorado SB 26-189 signedUnited StatesReplaces the 2024 AI Act with a notice law from Jan 2027Source
Jul 24, 2026EU Digital Omnibus publishedEUHigh-risk deadlines moved; ban on AI intimate imagery from Dec 2, 2026Source
Sep 2026UK still without an AI billUnited KingdomNo bill introduced; next King's Speech not expected before late 2027Source

Registry

Four registry incidents drew regulator or court action relevant to regulatory frameworks

Registry incidents linked to this page
IncidentDateRegulator or court action
Cruise Robotaxi Drags a PedestrianOct 2023California DMV suspended Cruise's driverless permits on October 24, 2023. NHTSA fined Cruise $1.5 million on September 30, 2024, for incomplete crash reports. On November 14, 2024, Cruise admitted filing a false report and agreed to a $500,000 deferred prosecution agreement.
Tesla Recalls Autopilot After NHTSA InvestigationDec 2023NHTSA opened PE21-020 in August 2021, upgraded it to EA22-002 in June 2022 and closed it in April 2024 after reviewing 956 crashes, opening RQ24-009 on whether the fix worked. No penalty.
Xiaomi SU7 Crashes on Assisted DrivingMar 2025In April 2025 MIIT barred carmakers from calling driver assistance autonomous driving and required approval of driving-function updates. On September 19, 2025, Xiaomi recalled 116,887 SU7s through SAMR over a highway driver assistance flaw.
Waymo Robotaxis Pass Stopped School BusesSep 2025NHTSA opened preliminary evaluation PE25-013 on October 17, 2025, and wrote to Waymo again on December 3. Waymo counts 12 citations tied to the behavior.

Rules

Three models: one horizontal law (EU, Korea), layered sector rules (China), federal policy plus state laws (US)

Key instruments: regulatory frameworks
RuleJurisdictionWhat it requiresStatus
EU AI Act as amended by 2026/1744EUProhibitions and general-purpose duties live; high-risk delayedIn force
China's layered AI regimeChinaAlgorithm, deep synthesis, generative AI, labeling and agent rulesIn force 2022 to 2026
America's AI Action Plan and EO 14179US federalPromotion-first federal policy; Biden order revokedJul 2025
Korea AI Basic ActSouth KoreaHigh-impact AI duties and labelingEffective Jan 22, 2026
Japan AI Promotion ActJapanPromotion law with guidance, no penaltiesEnacted 2025
Texas TRAIGATexasBans AI built to incite harm or discriminate; 60-day cureEffective Jan 1, 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); registry incidents (4 linked)
Key sources
China agent Implementation Opinions
Korea AI Basic Act takes effect
Colorado SB 26-189 signed
EU Digital Omnibus published
UK still without an AI bill
Last verified
October 2026