Comparison Tables

Side-by-side comparison matrices of AI governance approaches across major jurisdictions — regulatory approach, penalties, enforcement, individual rights, and sectoral coverage.

Last updated: February 2026 6 Matrices 10+ Jurisdictions

Table of Contents

1. Regulatory Approach Comparison

Jurisdiction Approach Primary Instrument AI Definition Risk Classification
EUComprehensive horizontal regulationAI Act (Regulation 2024/1689)Broad — machine-based system with autonomy, inference capability4 tiers: Unacceptable, High, Limited, Minimal
US (Federal)Sector-specific; executive action; voluntaryExecutive Orders; agency guidance; NIST RMFVaries by agency; no single federal definitionNo single classification; sector-dependent
US (States)Fragmented; issue-specific lawsState laws (Colorado, Illinois, NYC, etc.)Varies by stateColorado: “high-risk” for consequential decisions
ChinaTechnology-specific regulationsAlgorithm, Deep Synthesis, GenAI rulesTechnology-specific definitions per regulationBy technology type, not risk level
UKPro-innovation; sector regulatorsWhite Paper; existing regulators (FCA, Ofcom, etc.)Adaptability and autonomy-focusedCross-sectoral principles; sector risk assessment
CanadaProposed legislation + existing directivesAIDA (proposed); ADM Directive (federal gov)AIDA: technological system using ML or similarADM Directive: 4-level impact assessment
JapanSocial Principles; soft law; sector guidanceSocial Principles of Human-Centric AI (2019)Broad; not formally defined in binding lawNo formal classification
South KoreaComprehensive framework + sector lawsAI Framework Act (2025)Defined in framework lawHigh-risk AI designation
SingaporeGovernance framework; voluntaryModel AI Governance FrameworkBroad; technology-neutralRisk-proportionate; sector-based
BrazilProposed comprehensive lawAI Bill (PL 2338/2023)Broad; based on OECD definitionRisk-based (influenced by EU AI Act)

2. Penalties & Enforcement

Jurisdiction Maximum Penalty Enforcement Body Private Right of Action Criminal Liability
EU AI Act€35M or 7% global turnover (prohibited practices)National authorities + AI Office (GPAI)Not directly; through national lawNot in AI Act; possible under national law
EU GDPR€20M or 4% global turnoverNational DPAs (Data Protection Authorities)Yes (Articles 79, 82)Varies by member state
US FederalVaries by agency (FTC: unlimited in theory)FTC, EEOC, HHS, SEC, CFPB, etc.Agency-specific; tort law availableOnly for specific violations (fraud, etc.)
Illinois BIPA$5,000 per intentional violationState AG + private litigationYes (major source of class actions)No
Colorado AI Act$20,000 per violation (UCPA penalties)State Attorney General (exclusive)NoNo
ChinaUp to 10× illegal gains; license revocationCAC; sector regulatorsLimitedPossible for serious violations
UKSector-dependent (e.g., ICO: £17.5M/4%)Sector regulators (ICO, FCA, Ofcom, etc.)Sector-dependentSector-dependent
Canada (proposed)C$25M or 5% global revenue (AIDA)AI & Data Commissioner (proposed)Proposed in AIDAYes for reckless/serious harm (proposed)
South KoreaUp to 3% of relevant revenueAI Committee; sector agenciesYes for damages from AINo specific AI criminal provisions
Brazil (proposed)2% of revenue (up to R$50M per violation)National AI Authority (proposed)Yes (proposed)Under discussion

3. Individual Rights

Right EU US China UK Canada
Right to know AI is being used AI Act Art. 50 Varies by state/sector Multiple regulations ICO guidance ADM Directive
Right to explanation of AI decision GDPR Art. 22; AI Act Art. 86 ECOA (credit); no general right Algorithm rules require disclosure UK GDPR Art. 22 ADM Directive
Right to human review GDPR Art. 22(3); AI Act Art. 14 Sector-specific only Limited provisions UK GDPR Art. 22 ADM Directive
Right to opt out of AI profiling GDPR Art. 21 State privacy laws (CA, CO, etc.) Algorithm recommendation rules UK GDPR Art. 21 PIPEDA principles
Right to contest AI decision AI Act Art. 86 Due process (gov only) Complaint mechanisms No specific right ADM Directive
Right to non-discrimination by AI AI Act; equality directives Civil Rights Act; EEOC General equality provisions Equality Act 2010 Human Rights Act

= Explicit legal provision   = Partial or indirect coverage

4. Scope & Definitions

Dimension EU AI Act US (EO 14110) China OECD
AI DefinitionMachine-based system designed to operate with varying levels of autonomy, that may exhibit adaptiveness, and that infers how to generate outputsMachine-based system that can, for a given set of human-defined objectives, make predictions, recommendations, or decisionsTechnology-specific (algorithm recommendation, deep synthesis, generative AI defined separately)Machine-based system that can, for explicit or implicit objectives, infer how to generate outputs such as predictions, recommendations, or decisions
Covers public sectorYes (with law enforcement provisions)Yes (primary focus of EO)YesYes
Covers private sectorYesLimited (agency-specific)YesYes (recommendations)
Extraterritorial reachYes (Brussels Effect — applies to systems used in EU)Limited (dual-use export controls)Yes (for services to Chinese users)N/A (non-binding)
Covers open-sourceLimited exemptions for open-source GPAINo specific provisionsYes (registration required)Addressed in 2024 update

5. Sector-Specific Regulation

Sector EU US China UK
HealthcareAI Act (high-risk) + MDR + IVDRFDA (SaMD framework; 950+ clearances)NMPA regulationsMHRA; NICE ESF
Financial ServicesAI Act + existing EU financial regulationOCC SR 11-7; SEC guidance; CFPBPBOC; CBIRC guidanceFCA; PRA SS1/23
EmploymentAI Act (high-risk) + Platform Workers DirectiveEEOC; NYC LL144; state lawsLabor law provisionsEquality Act; ICO guidance
Autonomous VehiclesType approval; AI ActNHTSA; state laws (50 states vary)MIIT regulations; city pilotsAutomated Vehicles Act 2024
Law EnforcementAI Act (high-risk + biometric prohibitions)No federal law; DOJ guidance; city bansExtensive deployment; limited public regulationSector-specific guidance; Bridges v. South Wales Police
EducationAI Act (high-risk for admissions, assessment)FERPA; ED guidanceMOE guidance; content requirementsDfE AI strategy
MilitaryNational security exemption in AI ActDoD 3000.09; NSMCMC guidelines; civil-military fusionMOD AI Strategy; JSP 936

6. Implementation Status (as of February 2026)

Jurisdiction Primary AI Law Status Key Dates Enforcement Active
EUAI ActPhased implementationProhibitions: Feb 2025; GPAI: Aug 2025; High-risk: Aug 2026Partial (prohibitions enforced)
US FederalNo comprehensive lawExecutive action; agency rulesOngoing (EO 14110 revoked Jan 2025; agency actions continue)Yes (existing authorities)
ColoradoSB 24-205Enacted; pre-enforcementEffective: Feb 2026Pending
ChinaMultiple regulationsFully operationalAlgorithm: Mar 2022; Deep Synthesis: Jan 2023; GenAI: Aug 2023Yes
UKNo comprehensive lawSector-based approachOngoing regulatory updatesYes (existing regulators)
CanadaAIDA (proposed)Stalled in ParliamentUncertainNo (ADM Directive: federal gov only)
South KoreaAI Framework ActEnacted 2025Enforcement beginning 2025-2026Yes (phased)
BrazilAI Bill (PL 2338/2023)Under parliamentary reviewExpected 2026No
JapanNo binding AI lawGuidelines-basedOngoing updates to guidelinesSoft enforcement via guidance
SingaporeNo binding AI lawGovernance framework + PDPAFramework updated periodicallyPDPA enforced; AI framework voluntary
Previous Timeline Next Home