Scientists Called for the Creation of an International Body to Oversee Advanced AI

• About 400 scientists have called for the introduction of international isolation standards for advanced AI systems and the creation of a security incident registry.
• Researchers believe that models with cyber capabilities should, at an early stage, be fully isolated from external networks and placed under independent oversight.

About 400 scientists from 34 countries have called for the creation of an international body to oversee the safety of advanced artificial intelligence (AI) systems. Signatories of the open letter include field pioneers Geoffrey Hinton and Yoshua Bengio, Reuters reported.

The proposed organization would set industry-wide standards for model isolation and maintain a registry of significant security incidents. In the researchers’ view, this registry should eventually become public.

Researchers call for stronger model isolation

In the open letter, scientists said that advanced AI models with cyber capabilities should, at an early stage, be fully isolated from external networks using air gap technology. They also called for independent oversight and stressed that voluntary commitments by tech companies are not enough.

The signatories believe that organizations developing or testing such models should immediately implement high-assurance isolation standards. This needs to be done while governments work on creating a more formal oversight framework.

The letter was also a response to a series of incidents that, according to the researchers, exposed weaknesses in the testing and isolation of advanced AI systems. In particular, it referred to cases where models bypassed insufficiently secured “sandboxes” and compromised external systems.

Governments debate approaches to AI regulation

Journalists noted that policymakers continue to debate how to control advanced artificial intelligence systems. EU tech chief Henna Virkkunen told Reuters on Friday that the European Union’s Artificial Intelligence Act (AI Act) is sufficient to address the risks associated with uncontrolled AI systems.

US President Donald Trump’s administration lifted a block that had prevented American labs from giving the UK AI Safety Institute early access to models to assess their safety.

As a reminder, earlier media reports said that OpenAI and Anthropic began preparing for a potential AI catastrophe and its consequences.

Сообщение Scientists Called for the Creation of an International Body to Oversee Advanced AI появились сначала на INCRYPTED.