AI Experts Urge Urgent Safeguards Against Rogue Artificial Intelligence

ISLAMABAD: (Web Desk) More than 200 artificial intelligence and cybersecurity researchers have called for urgent measures to prevent AI models from escaping containment and threatening critical infrastructure.

In an open letter issued on Tuesday, the researchers urged organisations to physically disconnect AI models capable of conducting cyberattacks from the internet during testing. The signatories include Nobel laureate Geoffrey Hinton and Turing Award winner Yoshua Bengio, two prominent figures in artificial intelligence research.

The researchers stressed the need for stronger AI safety research alongside more effective containment technologies and procedures.

Their warning follows disclosures by leading US-based AI companies in recent months about incidents in which models under development reportedly bypassed containment measures, accessed the internet and attempted to hack websites. Some incidents involved unauthorised access to Australian government websites.

“Without immediate changes in AI development and deployment practices, such incidents will continue to occur, and future incidents will be far more damaging to our exposed critical infrastructure,” the researchers warned.

The letter called for organisations developing, testing or deploying cyber-capable AI models to use air-gapped systems, which are physically isolated from the internet, to reduce the risk of unauthorised access or escape from controlled environments.

The signatories also proposed establishing an international body to set common security standards and best practices. They said organisations should be required to report containment failures to the proposed body, which would maintain records of such incidents.

“This transparency is required for both learning and accountability purposes,” the letter stated.

The incidents have intensified calls for tighter oversight of AI development. However, governments and technology companies remain divided over how far development should be restricted and what form regulation should take.

In the United States, the Trump administration has opposed binding regulations that it believes could slow innovation amid growing technological competition with China. Instead, it has promoted a voluntary code of conduct.

The United States currently has no federal law specifically governing these AI models.