GLOBAL AI POLICY ANALYSIS • 2025–2026

AI Governance in a Fragmented Regulatory Landscape

Mapping jurisdictions by regulatory philosophy and governance maturity

Guidance & Principles

Voluntary guidance, principles-based approaches and non-binding frameworks

Governance & Assurance Frameworks

Established governance frameworks, testing, standards, assurance mechanisms and implementation guidance

Binding Legal Requirements

Enforceable laws and regulations with binding obligations and penalties

RISK-MANAGEMENT POSTURE

Anticipatory / Preventive

Focus on preventing harm and addressing risks early

Balanced / Adaptive

Seeks proportionate risk management as capabilities evolve

Reactive / Emerging / Fragmented

Responds to risks or is at an early stage of regulation

♞

Singapore

✦

European Union

▥

China

✣

South Korea

⌁

Australia

✚

United Kingdom

✦

Canada

▰

Japan

✣

Saudi Arabia

▥

India

✚

Brazil

●

US Federal
Agencies, EOs, Guidance

▤

US States
CA, CO, UT...

LOW REGULATORY PRESSURELower legal pressure
Less prescriptive

HIGH REGULATORY PRESSUREHigher legal pressure
More prescriptive

REGULATORY PRESSURE

JURISDICTIONS: STATUS & APPROACH

Hover over a jurisdiction on the map to highlight the relevant description here.

✦

European Union

AI Act in force; comprehensive risk-based regime with binding duties, conformity controls and penalties.

▥

China

State-led binding rules for recommendation algorithms, deep synthesis and generative AI services.

✣

South Korea

AI Basic Act establishes national governance architecture with trust, safety and oversight obligations.

♞

Singapore

Model AI Governance Framework, AI Verify, testing and assurance tools; mature but largely voluntary.

⌁

Australia

Voluntary AI safety standard plus proposed mandatory guardrails for high-risk AI; moving upward toward preventive controls.

✦

Canada

AIDA / Bill C-27 failed; no federal AI Act currently. Risk-based policy direction remains, but less advanced than Australia.

✚

United Kingdom

No comprehensive AI Act. Existing laws enforced by FCA, ICO, CMA and PRA. Strong governance ecosystem, including AI Safety Institute, guidance and model assurance.

▰

Japan

AI Act and guidelines create a lighter, promotion-led framework; more adaptive than EU-style ex ante control.

✣

Saudi Arabia

National AI Strategy and AI Ethics Principles; strategy-led governance rather than comprehensive binding law.

▤

US States

Patchwork of state AI laws and bills, including transparency, high-risk AI and sector-specific duties.

●

US Federal

No comprehensive federal AI law; executive orders, agency guidance and enforcement through existing powers.

▥

India

No comprehensive AI law; advisory principles and sectoral or platform-specific guidance remain emerging.

✚

Brazil

AI Bill 2338/2023 remains in development; risk-based framework influenced by the EU approach.

How to read it: jurisdictions are mapped by regulatory philosophy (vertical) and regulatory pressure (horizontal); placement reflects governance maturity and implementation architecture rather than solely the binding nature of obligations. Hover or tap any jurisdiction for its status and approach.

Source: “AI Governance in a Fragmented Regulatory Landscape” • IFK BioLiterate AI Ethics & Governance Article 1
Note: Placement reflects governance maturity and implementation architecture rather than solely the legal binding nature of obligations.

Despite divergent approaches, jurisdictions converge around core governance expectations: accountability, transparency, risk management, controllability and auditability.