INTERNATIONAL · 9 MIN READ

Global AI Regulation Is Becoming a Board-Level Tracking Problem.

Multinational organisations face a growing challenge as AI requirements develop differently across jurisdictions, increasing the need for coordinated regulatory intelligence and enterprise-wide governance.

INTAI Research · Published 2 June 2026

Read the Update

Divergence, Not Convergence, Is the Current Trajectory

For a period, it was plausible to expect major jurisdictions to converge on broadly similar AI regulatory approaches, following the pattern data protection law took after GDPR. That expectation has not held. The current global picture is one of genuine, structural divergence across legal traditions, priorities and enforcement philosophies.

The EU has continued down the path of binding, horizontal regulation with extraterritorial reach, the AI Act applies to organisations serving EU customers regardless of where those organisations are headquartered, with GPAI enforcement powers and penalties of up to €35 million or 7% of global turnover activating from 2 August 2026. The United States has moved in a different direction: on 11 December 2025, the federal executive branch signed “Ensuring a National Policy Framework for Artificial Intelligence,” signalling an intent to consolidate AI oversight federally and counter the rapid proliferation of individual state-level AI laws, even as Congress separately moves toward more targeted children’s safety legislation.

Three Distinct Models, Operating Simultaneously

China has taken a third path entirely: a dual-drive governance model combining binding national law with detailed technical standards, producing more sector-specific AI regulation between 2021 and 2025 than any other jurisdiction in the world, and embedding AI governance directly into national law through Cybersecurity Law amendments that took effect on 1 January 2026.

Smaller, more agile jurisdictions are adding further texture to the picture: Singapore launched what is described as the world’s first agentic AI governance framework in January 2026, requiring verifiable digital identity and audit trails for autonomous agents specifically, a level of technical specificity neither the EU nor the US frameworks currently match. For a multinational organisation, this is not three or four variations on a similar theme. It is genuinely different regulatory philosophies, running concurrently, each demanding a different kind of evidence.

Central Insight

Multinational organisations cannot build a single, unified AI compliance programme and expect it to satisfy every jurisdiction they operate in, that approach assumes a convergence that has not materialised and shows little sign of arriving soon. The realistic alternative is a core governance framework, built around the most stringent common requirements, with jurisdiction-specific compliance layers added where local law genuinely diverges, rather than attempting to average across fundamentally different regulatory philosophies.

Why This Has Become a Board-Level Problem, Not Only a Legal One

The pace and divergence of this landscape mean that AI regulatory tracking can no longer sit solely inside a legal or compliance function that reviews it periodically. It requires the same kind of continuous, coordinated intelligence function that multinational organisations already run for tax, trade and sanctions exposure, because a regulatory change in one jurisdiction, such as the EU’s May 2026 high-risk deferral or China’s January 2026 Cybersecurity Law amendment, can materially change an organisation’s compliance obligations with very little advance notice.

For boards overseeing organisations with meaningful cross-border AI exposure, the practical question worth asking directly is not “are we compliant with AI regulation”, a question that has no single answer across a fragmented global landscape, but “do we have a functioning process for knowing when a relevant jurisdiction's requirements change, and how quickly can we assess what that change means for us.” That tracking capability is, increasingly, the actual governance deliverable.