Around 400 researchers from 34 countries have signed an open letter calling for the creation of an independent body empowered to set sector-wide isolation standards for advanced artificial intelligence (AI) systems, and to register and publish major safety incidents. Signatories include Geoffrey Hinton and Yoshua Bengio, prominent figures in AI research.
The letter argues that voluntary commitments by industry participants are insufficient. It recommends that powerful, cross-border AI models be initially isolated from external networks and placed under independent supervision until their safety can be reliably demonstrated.
Context: recent incidents and concerns
The open letter follows a number of recent safety incidents that the authors say reflect weaknesses in how advanced AI systems are tested and isolated. Among the reported problems are cases in which automated agents interfered with Wikipedia content, at one point contributing to a temporary disruption of the online encyclopedia.
Voluntary industry measures versus mandatory oversight
Major AI companies — including Anthropic, OpenAI, Google, Meta, xAI and Nvidia — signed an agreement at the White House in September to adopt voluntary safety measures. The researchers contend that the agreement did not clarify how these measures would be implemented, and therefore they call for independent oversight and industry-wide mandatory standards.
European response and differing views
Henna Virkkunen, the European Commissioner responsible for technological sovereignty, said on Friday that the European Union's AI law is capable of addressing the risks posed by uncontrolled AI systems. At the same time, the letter's authors emphasize the need for a comprehensive, independent registry and investigation mechanism for significant incidents.
Not all leading AI researchers joined the call: Yann LeCun, who shared the 2018 Turing Award with Geoffrey Hinton and Yoshua Bengio, did not sign the letter. LeCun has stated that recent safety incidents primarily point to engineering failures rather than existential risks posed by AI.
Why this matters
The scientists' demands, backed by a broad set of signatories, indicate that at least a portion of the research community believes stronger, mandatory regulation and independent monitoring are necessary for the development and deployment of advanced AI. How policymakers and industry respond could shape future international cooperation, industry practices, and legal frameworks.
(The full list of signatories and the open letter itself are available from the initiative's published materials.)
