Definition
Governments trying to agree on shared rules and safety standards for AI, so it’s governed consistently across borders instead of country by country.
At a glance
- Mostly voluntary summits and declarations, not binding treaties.
- The Bletchley Declaration (2023) drew 28 countries plus the EU — including the US and China[2].
- The UN now runs the first AI bodies covering all 193 member states[1].
- The practical upshot for you: AI rules differ by country, so compliance is not one-size-fits-all.
How it happens
AI crosses borders, so its risks do too. Coordination is a patchwork: AI Safety Summits (Bletchley 2023, Seoul 2024, Paris 2025), shared safety-testing institutes, and standards bodies. Seoul added voluntary commitments from 16 AI firms[4]. The goal is “interoperability” — rules that fit together well enough that companies aren’t stuck with contradictory regimes.
Why it stalls
Geopolitics. At Paris 2025, the US and UK refused to sign a statement 61 countries backed[3]. Underneath sits US-China rivalry and a split between Europe’s heavy regulation and America’s lighter touch[5]. With no global enforcer, agreements stay commitments, not law.
Bottom line
Coordination is real but loose — plan for AI compliance country by country, not one global standard.
References
- Secretary-General Welcomes General Assembly Decision to Establish New Mechanisms Promoting International Cooperation on Governance of Artificial Intelligence. United Nations press.un.org
- AI Safety Summit. Wikipedia en.wikipedia.org
- As US and UK refuse to sign the Paris AI Action Summit statement, other countries commit to developing open, inclusive, ethical AI. TechCrunch techcrunch.com
- AI Seoul Summit: 16 AI firms make voluntary safety commitments. Computer Weekly www.computerweekly.com
- Strengthening international cooperation on AI. Brookings Institution www.brookings.edu
Comments
Questions, corrections, and links welcome. Be specific and civil.