Global AI Regulation Tracker

This reference tracks the current state of AI regulation across major jurisdictions. Regulatory landscapes evolve rapidly; always verify current status with official sources before making compliance decisions. Last reviewed: April 2026.

Status Overview

JurisdictionRegulationStatusEffective Date
European UnionEU Artificial Intelligence Act (Regulation 2024/1689)EnactedAugust 1, 2024 (phased through August 2027)
United StatesExecutive Order 14110 on Safe, Secure, and Trustworthy AISigned October 30, 2023Various deadlines from 2024 onward; status depends on administration policy
ChinaMultiple AI Regulations (Algorithmic Recommendation, Deep Synthesis, Generative AI)EnactedAlgorithmic Recommendation: March 2022. Deep Synthesis: January 2023. Generative AI: August 2023.
United KingdomPro-Innovation Approach to AI RegulationFramework publishedNo single effective date; existing regulators implement AI principles within their mandates
CanadaArtificial Intelligence and Data Act (AIDA) -- Part of Bill C-27ProposedNot yet enacted; would require reintroduction in a new Parliament
BrazilBrazilian AI Bill (PL 2338/2023)Under legislative considerationNot yet enacted; effective date would be set upon passage

European Union

EU Artificial Intelligence Act (Regulation 2024/1689)

Status

Enacted

Scope

Providers, deployers, importers, and distributors of AI systems in the EU market, regardless of establishment location. Extraterritorial application to AI systems whose output is used in the EU.

Key Requirements

Risk-based classification (unacceptable, high, limited, minimal). Prohibited practices (social scoring, real-time biometric ID). Conformity assessment for high-risk systems. Transparency obligations. GPAI model requirements. Penalties up to 35M EUR or 7% global turnover.

Enforcement

EU AI Office; national market surveillance authorities in Member States.

United States

Executive Order 14110 on Safe, Secure, and Trustworthy AI

Status

Signed October 30, 2023 (portions rescinded by EO 14179, January 2025)

Scope

Federal agencies; companies developing dual-use foundation models above compute thresholds; federal contractors. Voluntary for private sector beyond federal procurement.

Key Requirements

Safety testing and reporting for dual-use foundation models. Red-teaming requirements. Watermarking and content authentication standards (NIST). Federal agency AI use case inventories. AI risk management aligned with NIST AI RMF. Note: EO 14179 (January 2025) rescinded certain provisions; verify current status.

Enforcement

NIST; DHS; DOE; individual federal agencies. No single federal AI regulator.

China

Multiple AI Regulations (Algorithmic Recommendation, Deep Synthesis, Generative AI)

Status

Enacted (incremental regulations since 2022)

Scope

AI service providers operating in China. Algorithmic recommendation services, deep synthesis (deepfake) providers, and generative AI service providers.

Key Requirements

Algorithm registration with the Cyberspace Administration of China (CAC). Content moderation and socialist core values alignment. User opt-out rights for algorithmic recommendations. Watermarking for AI-generated content. Security assessments before public deployment. Training data legality requirements.

Enforcement

Cyberspace Administration of China (CAC); Ministry of Science and Technology (MOST).

United Kingdom

Pro-Innovation Approach to AI Regulation

Status

Framework published (March 2023); sector-specific implementation ongoing

Scope

All sectors regulated by existing UK regulators (FCA, Ofcom, CMA, ICO, MHRA, etc.). Cross-cutting principles applied through existing regulatory frameworks rather than new AI-specific legislation.

Key Requirements

Five cross-cutting principles: safety/security/robustness; transparency/explainability; fairness; accountability/governance; contestability/redress. Sector regulators interpret and enforce within their domains. No mandatory registration or conformity assessment (contrast with EU). AI Safety Institute conducts frontier model evaluations.

Enforcement

Existing sector regulators (FCA, ICO, Ofcom, CMA, etc.). AI Safety Institute (evaluation, not enforcement).

Canada

Artificial Intelligence and Data Act (AIDA) -- Part of Bill C-27

Status

Proposed (introduced June 2022; died on order paper when Parliament prorogued January 2025)

Scope

Would have applied to persons responsible for AI systems in the course of international or interprovincial trade and commerce. High-impact systems subject to additional obligations.

Key Requirements

As proposed: classification of high-impact AI systems. Risk assessments and mitigation measures. Transparency requirements. Prohibition on AI systems that cause serious harm through reckless or malicious use. Criminal penalties for certain violations. Note: AIDA has not been enacted; a new bill may differ substantially.

Enforcement

Would have established an AI and Data Commissioner. Current status: no federal AI-specific legislation in force.

Brazil

Brazilian AI Bill (PL 2338/2023)

Status

Under legislative consideration (approved by Senate committee, pending full Senate and Chamber votes)

Scope

AI systems developed or used in Brazil, or whose effects occur in Brazil. Risk-based approach similar to the EU AI Act.

Key Requirements

As proposed: risk classification (excessive, high, general). Prohibited AI practices (social scoring, subliminal manipulation). Impact assessments for high-risk systems. Transparency obligations. Algorithmic auditing rights. Establishment of a national AI authority (SIA). Penalties up to 2% of revenue or 50M BRL.

Enforcement

Proposed: Sistema Nacional de Regulacao e Governanca de Inteligencia Artificial (SIA) with coordination role; sectoral regulators retain enforcement within their domains.

Disclaimer

This tracker is provided for informational purposes only. AI regulation is evolving rapidly across all jurisdictions. The status, scope, and requirements described above reflect publicly available information as of the date noted and may not reflect the most current legal position. Organizations should consult qualified legal counsel in each relevant jurisdiction before making compliance decisions.