The Paris AI Action Summit and a governance split (2025)
In February 2025 Paris hosted the AI Action Summit. Some 60 countries signed a declaration on inclusive, sustainable AI, but the United States and United Kingdom refused, exposing a rift over how to govern the technology.
On February 10 and 11, 2025, the AI Action Summit met at the Grand Palais in Paris, co-chaired by French President Emmanuel Macron and Indian Prime Minister Narendra Modi. It was the third in a series that began with the 2023 AI Safety Summit at Bletchley Park and continued in Seoul in 2024, and it marked a clear change of mood: away from existential-risk warnings and toward growth, investment, and opportunity.
A declaration, minus two
Around 60 signatories, including France, China, India, Japan and Canada, signed a joint Statement on Inclusive and Sustainable Artificial Intelligence for People and the Planet. It called for AI that is open, transparent, ethical and safe, for narrowing the digital divide, and for avoiding excessive market concentration.
The notable absentees were the United States and the United Kingdom, both of which declined to sign. US Vice President JD Vance used his appearance to warn against “excessive” regulation, casting heavy rules as a brake on innovation. The UK said the text lacked practical clarity on global governance and did not do enough on national security.
Two theories of AI policy
The split was not really about the wording. It exposed two competing instincts: one that wants shared international guardrails, echoing the spirit of the EU AI Act, and one that treats regulation as a competitive disadvantage in a race for compute and models. The summit came just weeks after DeepSeek had shaken assumptions about who leads that race, which only sharpened the stakes.
Why it matters
Bletchley Park had produced a rare moment of consensus about AI safety. Paris showed how quickly that consensus frayed once money and geopolitics moved to the foreground. For a technology built across borders, the meeting was an early sign that the world may not agree on a single rulebook, and that how AI is governed could fragment along the same lines as who builds it.