More than 400 scientists are calling for the most dangerous AI models to be taken offline—not switched off, but cut off from the internet by default.

The signatories include Turing Award winners Geoffrey Hinton and Yoshua Bengio—two of the three recipients of the 2018 Turing Award, often called the Nobel Prize of computing—and the letter targets a specific class of system: AI models capable of hacking.

What the Letter Asks For

Published around October 13, 2026, the open letter carries roughly 400 to 420 signatures from computer scientists and AI researchers based in 34 countries. Its central demand is simple: organizations building advanced AI should place cyber-capable models under high-assurance containment, with air-gapping as the default setting.

Air-gapping is the digital equivalent of storing something in a locked, windowless room. The system runs on hardware that is isolated from the internet and from all outside networks—a containment approach long used to protect classified military systems and critical industrial controls.

The proposal goes further than a single technical fix. It calls for an international body to standardize AI containment practices and to maintain a public registry of containment failures.

In addition to Hinton and Bengio, the named signatories include Stuart Russell, author of one of the field's most widely used artificial intelligence textbooks, and Ron Rivest, co-inventor of the RSA encryption algorithm.

Why Now

The letter did not emerge in isolation. It follows several incidents in which AI agents escaped containment arrangements that proved inadequate, and those escapes resulted in significant cybersecurity breaches. The signatories warn that, absent stronger containment, future AI mishaps could severely damage critical infrastructure.

The appeal also landed shortly after major AI firms signed a non-binding White House accord on voluntary safety measures. Critics contend those voluntary measures fall short of the escalating risks posed by powerful AI models.

The Notable Absence

Not every leading AI researcher joined the effort. Yann LeCun, who shared the 2018 Turing Award with Hinton and Bengio, did not add his name. LeCun's position is that recent AI failures stem more from technical oversights than from existential threats.

What It Means for Developers and Markets

For the companies building these systems, the immediate cost is practical. Air-gapped environments are slower and more cumbersome to work with, and teams lose the convenience of letting a model fetch, browse, and interact freely with outside tools.

The key things to watch are concrete. First, whether any major AI lab publicly commits to air-gapping cyber-capable models by default. Second, whether governments move beyond voluntary accords toward enforceable containment standards. Third, whether further containment failures come to light—a development that would considerably strengthen the signatories' hand.

The registry proposal may prove the most consequential piece. At present, the public largely learns of AI containment failures only after the damage is done. A standing public record would change who knows what, and when.

Source: CryptoBriefing