Call for an International Regulatory Body for Artificial Intelligence Systems

Nearly 400 scientists, including Geoffrey Hinton and Yoshua Bengio, have published an open letter calling for the establishment of an international oversight body to ensure the security of advanced artificial intelligence systems.

International Oversight and Security Standards

The open letter, signed by researchers from 34 different countries, stated that common security standards must be established, particularly for artificial intelligence models possessing cyberattack capabilities.

It was requested that the proposed international institution determine industry-wide control standards, record serious security breaches, and make these records public.

Isolation Methods and Independent Auditing

Scientists argued that advanced artificial intelligence models should initially be operated in systems physically or logically isolated from external networks.

Through this method, known in the field of cybersecurity as air-gapping, it was proposed that systems be severed from direct connection to external networks and subjected to independent auditing.

Corporate Commitments and the Role of Governments

The letter explicitly emphasized that the security commitments voluntarily announced by companies remain insufficient against current risks.

Until governments establish more comprehensive oversight mechanisms, organizations developing artificial intelligence were demanded to urgently implement high-assurance standards.