A leaked draft of the Paris AI summit declaration has sparked significant criticism due to its lack of substantial progress on safety commitments, with experts raising alarms over its failure to address key issues outlined at the Bletchley Park summit in 2023. The Paris summit, which begins with world leaders and AI company heads gathering in the French capital on Monday, has moved away from its previous focus on safety concerns, according to Professor Stuart Russell, president of the International Association for Safe and Ethical AI. The summit, co-hosted by French President Emmanuel Macron and Indian Prime Minister Narendra Modi, includes prominent figures such as U.S. Vice President JD Vance, German Chancellor Olaf Scholz, European Commission President Ursula von der Leyen, and a number of leaders from major AI companies, including OpenAI’s CEO Sam Altman and Google’s Sundar Pichai.
Despite the high-profile attendance, Russell expressed concern about the lack of meaningful commitments in the draft declaration, specifically pointing out that a section pledging safety through an “open, multi-stakeholder and inclusive approach” seemed vague and ineffective. Of particular concern was the omission of commitments from the Bletchley Park declaration, which focused on addressing risks related to AI in areas such as cybersecurity, biotechnology, and disinformation. Russell warned that if this omission persisted in the final statement, it would represent “negligence of an unprecedented magnitude.” He expressed hope that this was simply a drafting error and that the final statement would reassert the urgency of addressing these critical risks.
The leaked draft also highlighted some geopolitical tensions, particularly with the U.S. government, which has reportedly pushed to exclude any mention of AI’s environmental impact, existential risks, or the role of the United Nations in AI governance. Macron’s representative, Anne Bouverot, acknowledged the difficulties in reaching consensus on these issues, emphasizing that the summit would proceed with a “coalition of the willing” approach. This reflects ongoing challenges in aligning global positions on AI safety and regulation.
On the agenda for the summit are two major outcomes: a final communique and the announcement of a new public interest foundation for AI, aimed at promoting the responsible development and use of AI. Macron is expected to announce the creation of this foundation, which will be funded by a combination of government, business, and philanthropic contributions. It will focus on collecting large datasets for public bodies and supporting AI projects that serve the public good. Mike Bracken, founder of Public Digital, emphasized the significance of this initiative, calling it a necessary step toward positioning AI as a public asset. He noted, however, that major U.S. AI companies may be reluctant to participate in the foundation, with the exception of open-source platform Hugging Face.
The summit also serves as a platform for high-level AI industry discussions, with figures like Matt Clifford, AI advisor to the UK Prime Minister, noting the growing importance of the summit in fostering global AI collaboration. Meanwhile, Sir Demis Hassabis, CEO of Google DeepMind and a Nobel prize winner, underscored the need for international cooperation on AI regulation, acknowledging the difficulties of achieving consensus in the current political climate. He also highlighted the competitive nature of AI advancements, specifically noting that while Chinese systems like DeepSeek are impressive, they do not represent groundbreaking scientific progress compared to Google’s latest models.
In parallel with the summit, the UK government is launching bidding for new AI growth zones, where data centers can be quickly established in areas with strong grid connections. The first confirmed site is in Culham, Oxfordshire, near the UK Atomic Energy Authority, with additional potential sites in Warrington and North Yorkshire.
The summit, co-hosted by Macron and Modi, marks a critical moment in the ongoing global debate over AI safety, governance, and regulation, but experts, such as Professor Russell, remain concerned that the lack of meaningful commitments could undermine efforts to address AI’s potential risks to society.
