Leaked draft: the statement set to be signed at the Paris AI Action Summit barely mentions AI risks and fails to follow up on commitments from previous Summits
an omission which would represent a huge blow to the Summit's French organizers. [image] Shakeel / @shakeelhashim : Other experts criticised the statement too. @_andreamiotti called it “a wasted opportunity to build upon the historic progress made at the Bletchley and Seoul Al Safety Summits” Another expert said the draft in its current form would be “inexcusably derelict” Shakeel / @shakeelhashim : And it really *does* fail to follow up on commitments. In Seoul, countries promised to develop risk thresholds and capability red-lines. Those aren't mentioned *at all* in the Paris statement. [image] Shakeel / @shakeelhashim : The statement, titled “Statement on Inclusive and Sustainable Artificial Intelligence for People and the Planet,” barely mentions Al risks and completely fails to follow up on previous summit commitments. .@herbiebradley said it “says effectively nothing except for platitudes”. [image]
Context & Ripple Effects
The Paris draft marks a sharp change from the Bletchley Declaration’s broad international safety alignment and the Seoul process’s stated work on risk thresholds and capability red-lines. It also echoes an earlier warning sign: a leaked UK summit communiqué had already favored existing collaborations over creating new risk-research machinery.
The omission matters because the summit’s joint statement is a vehicle for turning political attention into shared next steps. Related coverage indicates Paris was increasingly framed around investment and competition, while later disagreement over the final international agreement exposed how fragile consensus on “safe” AI had become.
First-order effects
- If retained in the final text, the draft leaves French organizers without a clear summit-level follow-through mechanism for the Seoul commitments on thresholds and red-lines.
- Safety-focused participants lose a prominent multilateral forum for reaffirming those specific commitments, while the Paris agenda shifts its public emphasis away from AI-risk management.
Second-order effects
- Countries and institutions that want operational risk rules may need to pursue them through separate national, regional, or smaller-group processes rather than rely on a Paris-led consensus.
- The gap strengthens the practical contrast between investment-and-competition priorities and governance proposals centered on measurable capability limits and risk thresholds.
Third-order effects
- If successive summits substitute general principles for implementation commitments, global AI governance could become more fragmented: declarations may sustain diplomatic alignment while concrete assurance rules emerge unevenly elsewhere.
- The episode points to a broader contest over whether international AI forums primarily coordinate industrial strategy or build durable mechanisms for managing advanced-system risks.
The trend: AI summit diplomacy is moving from broad safety consensus toward a harder trade-off between competitiveness, sovereignty, and operational risk governance.