About 400 scientists from 34 countries called for an international body to oversee advanced artificial intelligence systems in an open letter reported by Reuters on Oct. 13, 2026. The signatories, including AI pioneers Geoffrey Hinton and Yoshua Bengio, said common isolation standards and independent oversight are needed because voluntary commitments by technology companies are insufficient.

The proposed organization would establish industry-wide standards for isolating advanced AI models and maintain a registry of significant security incidents. The researchers said the registry should eventually become public.

Researchers seek stronger model isolation

The scientists said 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 of those systems.

Organizations developing or testing such models should immediately introduce high-assurance isolation standards while governments work to establish a more formal oversight framework, according to the letter.

The researchers said a series of incidents had exposed weaknesses in the testing and isolation of advanced AI systems. They cited cases in which models bypassed insufficiently secured sandboxes and compromised external systems.

Governments debate AI oversight

Policymakers continue to debate how to regulate advanced AI systems. EU tech chief Henna Virkkunen told Reuters on Friday that the European Union’s Artificial Intelligence Act was sufficient to address risks associated with uncontrolled AI systems.

U.S. President Donald Trump’s administration also lifted a block that had prevented American laboratories from giving the UK AI Safety Institute early access to models for safety assessments.

Earlier media reports said OpenAI and Anthropic had begun preparing for a potential AI catastrophe and its consequences.

Source: Incrypted