Global AI Regulation in 2025: A Comparative Overview

· 5 min read
ai compliance policy regulation technology

Note: This article represents a practitioner’s interpretation of the relevant rules and regulations in place at the time of writing. I am not a lawyer, and readers should consult with their own legal counsel and compliance teams before taking any action based on this information.

As artificial intelligence continues to transform our world, governments and regulatory bodies are racing to establish frameworks that ensure responsible AI development while fostering innovation.

This article introduces a new comprehensive series examining AI regulation across major jurisdictions, offering a glimpse into how different regions are approaching this critical challenge.

I’m hoping to offer a practitioner’s perspective on the key regulations and requirements in each jurisdiction, and practical approaches for building AI systems in the world we live in. I’m not overly interested in opineing on the merits of the regulations one way or another.

Over the coming weeks, I’ll publish detailed analyses of AI regulatory frameworks in eight key jurisdictions. Each deep dive will examine the unique aspects of that region’s approach, implementation requirements, and practical implications for organizations developing or deploying AI systems.

The European Union: Enforcement Begins

The EU’s AI Act is no longer just on the books; it’s being enforced. Prohibitions on unacceptable-risk AI took effect in February 2025, GPAI model obligations kicked in August 2025, and full high-risk system enforcement arrives in August 2026. A November 2025 Digital Omnibus proposal may simplify some requirements, but the regulatory machine is in motion.

Read the full EU AI Act analysis →

China: AI Enters National Law

China’s regulatory pace accelerated in 2025. New AI labeling rules took effect, the State Council issued an ambitious AI Plus Action Plan, and the legislature amended the Cybersecurity Law to incorporate AI provisions for the first time. More than 30 new standards are expected in 2026, with agentic AI emerging as a regulatory focus area.

Read the full China analysis →

United States: Federal-State Tension

The US landscape shifted significantly in 2025. The Trump administration revoked the Biden-era AI executive order and issued new orders prioritizing deregulation and challenging state AI laws. Meanwhile, California, Colorado, and Illinois all enacted enforceable AI legislation taking effect in early 2026. The collision between federal deregulation and state-level action is the defining dynamic heading into 2026.

Read the full US analysis →

United Kingdom: Pro-Innovation with Principles

Post-Brexit, the UK has developed a distinctive “pro-innovation” approach that emphasizes principles-based regulation over prescriptive rules. Their framework focuses on high-risk applications while maintaining flexibility for technological advancement.

Read the full UK analysis →

Japan: From Soft Law to Framework Legislation

Japan’s approach used to rely entirely on “soft law” guidelines, but in May 2025 the country enacted its first AI-specific legislation, the AI Promotion Act. It’s deliberately light-touch, establishing institutional structures without heavy penalties, but it marks a meaningful shift.

Read the full Japan analysis →

South Korea: The Second Comprehensive Framework

South Korea’s AI Basic Act took effect on January 22, 2026, making it the second jurisdiction after the EU to bring a comprehensive AI regulatory framework into force. Enforcement Decrees are being finalized to flesh out the implementation details.

Read the full South Korea analysis →

India: Moving Beyond Hands-Off

India still has no specific AI legislation, but 2025 brought notable movement. MeitY released official AI Governance Guidelines, a Private Member’s Bill proposed a statutory AI Ethics Committee, and the amended IT Rules 2026 introduced mandatory labeling of AI-generated content. The Digital India Act remains in development.

Read the full India analysis →

Key Themes Across Jurisdictions

Revisiting these themes a year later, the landscape has changed meaningfully:

  1. Comprehensive Legislation Is Now the Norm: The EU, South Korea, Japan, and multiple US states have all enacted AI-specific laws. China has folded AI into its national cybersecurity law. Only the UK and India lack binding comprehensive frameworks, and both are moving in that direction.

  2. Divergent Philosophies on Stringency: The EU and US states like California and Colorado represent the “duty of care” model with real penalties. Japan and the UK explicitly prioritize innovation with minimal compliance burdens. The US federal government is actively trying to preempt stricter state laws.

  3. AI Content Labeling Is Going Global: Both China (September 2025) and India (February 2026) have mandated labeling of AI-generated content. California’s TFAIA requires transparency reports. This is becoming a baseline expectation.

  4. Data Protection and AI Are Converging: The connection between AI regulation and data protection continues to tighten, as seen in the EU’s Digital Omnibus and India’s IT Rules amendments.

  5. Agentic AI Is the Next Frontier: China has already released draft rules addressing agentic AI systems, and other jurisdictions are watching closely. This will likely be a major regulatory theme through 2026 and beyond.

Looking Forward

As AI technology continues to evolve, these regulatory frameworks will undoubtedly adapt and change. Organizations must stay informed about requirements across jurisdictions while maintaining flexible compliance strategies that can evolve with the regulatory landscape.

The detailed analyses in this series will help practitioners understand the specific requirements, implementation strategies, and practical implications of each regulatory framework. Follow along as we examine how different regions are approaching the critical challenge of ensuring responsible AI development while fostering innovation.

Series Navigation

Changelog

  • February 2026: Updated all country summaries to reflect developments through early 2026. Revised Key Themes section to reflect the shift toward comprehensive AI legislation as the global norm.

Linked from