400 Researchers Call for an International Oversight Body for Artificial Intelligence

A large group of researchers, including two AI pioneers, called for the establishment of an international body to oversee and control this technology.

A large group of researchers, including two pioneers of artificial intelligence, called for the creation of an international body to monitor and control this technology.

According to Ambassador Arjan, quoted by Reuters, about 400 researchers, including Geoffrey Hinton and Yoshua Benjiu, pioneers of artificial intelligence, called for the formation of an international body to oversee containment standards and incident reporting requirements related to advanced artificial intelligence systems with cyber capabilities.

The proposed agency would create containment standards for the entire industry and maintain a repository of significant cyber incidents, which researchers say will gradually become public.

In an open letter, the researchers stated that advanced artificial intelligence models should initially be presented in isolation from external networks and be subject to independent oversight. They argue that the industry's voluntary commitments are not enough.

The proposed creation of such a body comes a month after Anthropic, OpenAI, Google, xAI, and Nvidia signed a White House agreement committing to implement voluntary safety measures without any formal enforcement mechanism.

According to the researchers, before presenting this letter, a chain of events occurred that revealed weaknesses in the way artificial intelligence systems were tested and controlled, including cases in which artificial intelligence models bypassed test environments and compromised external systems.

The signatories of the letter are from 34 countries around the world and believe that entities that develop or test cyber capability models should quickly implement high-assurance containment standards, and governments should move toward a more formal regulatory regime.

Policymakers are increasingly making arguments about how to monitor AI models. According to the EU's chief technology officer, Hanna Wirkoven, the "AI Act" is enough to address the risk of dangerous artificial intelligence systems.

This is while the American government has canceled the restrictions imposed on the British Artificial Intelligence Security Institute to access American artificial intelligence laboratories to test the safety of models.