Paris AI Action Summit Declaration (2025)

The Statement on Inclusive and Sustainable Artificial Intelligence for People and the Planet is a non-binding political declaration issued by participants of the AI Action Summit, co-chaired by France and India and signed on 10-11 February 2025. It was the third summit in the lineage that began with Bletchley and continued at Seoul, and it shifted the summit series' organising frame from AI "safety" toward "inclusive and sustainable" AI. The United States and the United Kingdom declined to sign.

Background

The Paris summit was the third in the series begun by the Bletchley AI Safety Summit and continued at Seoul. Hosted by France and India, the declaration moved away from the "safety" framing of Bletchley and Seoul toward an "inclusive and sustainable" framing, invoking the Paris Pact for People and the Planet and the UN Global Digital Compact. The text acknowledges prior international work including UN General Assembly resolutions, the Global Digital Compact, the UNESCO Recommendation on the Ethics of AI, the African Union Continental AI Strategy, OECD work, Council of Europe work, EU work, the G7 Hiroshima AI Process, and the G20. It commends the Bletchley and Seoul summits and notes their voluntary commitments.

Key provisions

The declaration sets out six priorities:

1. Promoting AI accessibility to reduce digital divides.
2. Ensuring AI is open, inclusive, transparent, ethical, safe, secure, and trustworthy.
3. Enabling AI innovation and avoiding market concentration.
4. Shaping the future of work and labour markets.
5. Making AI sustainable for people and the planet.
6. Reinforcing international cooperation.

It also announced several concrete deliverables. A Public Interest AI Platform and Incubator was established as a multi-stakeholder initiative for digital public goods, technical assistance, and capacity-building across data, models, compute, talent, openness, transparency, and audit. The summit held its first summit-level multi-stakeholder discussion on AI and energy, paired with an observatory on AI energy impact hosted with the International Energy Agency. The declaration created an observatories network on AI in the job market, intended to generate insights on workplace effects, training, productivity, and working conditions. It reaffirmed, per the UN Global Digital Compact, a Global Dialogue on AI Governance and an Independent International Scientific Panel on AI, the latter partially institutionalising the function of the International AI Safety Report.

Signatories

The declaration was signed by 64 parties: Armenia, Australia, Austria, Belgium, Brazil, Bulgaria, Cambodia, Canada, Chile, China, Colombia, Côte d'Ivoire, Croatia, Cyprus, Czech Republic, Denmark, Djibouti, Egypt, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, India, Indonesia, Ireland, Italy, Japan, Kazakhstan, Kenya, Latvia, Liechtenstein, Lithuania, Luxembourg, Malta, Mexico, Moldova, Monaco, Morocco, Netherlands, New Zealand, Nigeria, Norway, Peru, Poland, Portugal, Republic of Korea, Romania, Rwanda, Senegal, Serbia, Singapore, Slovakia, Slovenia, Spain, Sweden, Switzerland, Thailand, Ukraine, Uruguay, the African Union Commission, and the European Union.

The United States and the United Kingdom declined to sign. US Vice President JD Vance used his summit speech to signal a pivot away from safety framing, a stance consistent with Executive Order 14365 — Ensuring a National Policy Framework for AI and America's AI Action Plan. The United Kingdom cited concerns around national security language and governance provisions. The pattern broke up the alignment of the three states central to the Bletchley series: China signed the Paris declaration, while the US and UK did not.

Comparison with other approaches

Relative to the Bletchley Declaration, Paris de-emphasises "frontier AI" and "safety" as organising concepts and emphasises inclusion, sustainability, and public interest instead. The Seoul Commitments are corporate commitments, whereas Paris is inter-governmental; Paris only commends Seoul rather than directly building on it. Against America's AI Action Plan, the postures are oppositional: US non-signature at Paris reflects the Action Plan's innovation-first, deregulatory framing. Paris references the G7 Hiroshima Code of Conduct as acknowledged prior work but does not re-endorse its specific commitments.

Reactions and debates

Commentators identified several tensions in the declaration. On safety, the word survives in subordinate clauses but is no longer an organising principle; this de-emphasis sits in tension with the International AI Safety Report, whose function Paris nonetheless partially institutionalises via the Independent International Scientific Panel on AI. The US and UK refusal to sign was read as a withdrawal from the multilateral safety framework and the end of the Bletchley coalition structure. Critics also argued the six priorities are broad enough to command broad signature without binding signatories to specific action.

Observers noted a divergence in how China participated across the summit series: the Chinese government signed the Paris declaration but did not sign the Seoul Commitments as a state, where the signatories were Chinese frontier labs rather than the government.

Relationships

• supersedes (tonally): The Bletchley Declaration (AI Safety Summit, 1–2 November 2023) (as the summit-series organising document)
• depends-on: AI Race Dynamics, AI Safety Cases and Frameworks
• contradicts: Frontier AI Safety Commitments (Seoul, 2024) (in emphasis, not literal commitments)
• related: G7 Hiroshima Code of Conduct for Advanced AI (2023), America's AI Action Plan, Executive Order 14365 — Ensuring a National Policy Framework for AI, International AI Safety Report 2025, AI Environmental Impact