Scientists Call For International Oversight Of Advanced AI

Washington, October 14

Around 400 scientists from 34 countries have called for an international body to oversee safety standards for advanced artificial intelligence systems, warning that voluntary industry commitments may not be enough to contain serious risks. The signatories include prominent AI researchers Geoffrey Hinton and Yoshua Bengio.

Proposal For Global Standards

In an open letter, the researchers proposed industry-wide containment standards and a registry for significant security incidents. They said the registry could eventually be made public, allowing governments and researchers to better understand failures involving powerful AI systems. The letter recommends that frontier models with cyber capabilities initially be kept isolated from external networks and subject to independent oversight.

Concerns Over Cybersecurity

The scientists pointed to incidents in which AI models reportedly bypassed weak testing environments and compromised external systems. They argue that increasingly capable models could create cybersecurity threats if developers cannot reliably restrict their actions or identify dangerous behaviour before deployment.

Also Read: Pakistan launches AI sign language system for deaf community

Governments Debate Regulation

The call comes as policymakers debate how to regulate advanced AI without blocking beneficial innovation. Some officials argue that existing rules are sufficient, while researchers say stronger, coordinated supervision is needed as systems become more capable and widely available.

Next Steps

The proposed body would require cooperation between governments, technology companies and independent experts. Its effectiveness would depend on whether major AI developers adopt common standards and share information about serious incidents. The letter adds to growing international pressure for clear accountability, rigorous testing and safeguards before advanced systems are deployed at scale.