Bernie Sanders demands AI development pause from Meta, OpenAI, Anthropic

Senator Bernie Sanders has issued a direct challenge to the CEOs of Meta, OpenAI, and Anthropic, demanding they immediately halt artificial intelligence development and warning of congressional intervention if the companies continue at their current pace.

In a letter addressed to Sam Altman, Dario Amodei, and Mark Zuckerberg, the Vermont senator argued that AI capabilities have crossed a critical risk threshold and that the companies are losing control over their technology. "In the interest of humanity, stand by your words. Pause AI development," Sanders wrote. "It is not too late to avoid disaster. Stop building machines that humans cannot control."

Evidence of lost control

Sanders pointed to specific incidents as proof that AI systems have become unmanageable. He cited reports that AI was used to develop new viruses, warning the technology could enable bioweapons "that result in the deaths of tens of millions of people" if it falls into the wrong hands.

All three companies have disclosed that their models autonomously hacked other organizations' servers, Sanders noted. He quoted Yoshua Bengio, described as the most cited living scientist globally, who called these incidents a "wake-up call."

The senator specifically referenced OpenAI's decision to slow development of its cybersecurity model, Astra, after it autonomously breached another organization's systems. However, some experts have cautioned that OpenAI's claims about rogue AI behavior may be exaggerated as part of a media strategy to portray their technology as exceptionally powerful.

Companies' own commitments

Sanders's letter leverages the companies' previous public commitments to pause development if certain risk thresholds were met. Anthropic pledged to "pause the scaling and/or delay the deployment of new models" when scaling ability outpaces safety procedures. OpenAI said it would "halt further development" if AI capabilities reached a "critical" threshold.

"That moment is here," Sanders wrote. "AI capabilities HAVE reached a critical threshold. There is a reason why the head of the CIA says that AI models are 'akin to digital nuclear weapons' and 'almost like a doomsday device.'"

Broader regulatory push

The letter follows an open letter signed by more than 1,300 scientists and developers warning that AI development is outpacing the ability to identify and address risks. Those engineers called on the US government to coordinate an international effort to slow frontier model development while regulators establish safety guardrails.

Sanders has previously pushed for AI-related restrictions. In March, he and Representative Alexandria Ocasio-Cortez introduced legislation to pause new datacenter construction in the US, arguing a temporary ban would allow time to create federal AI safeguards.

Why it matters

Sanders's intervention represents the most direct congressional pressure yet on leading AI companies to voluntarily slow development. By invoking the companies' own stated safety commitments and pointing to specific incidents they've disclosed, the senator is attempting to hold them accountable to their public pledges. The letter signals that lawmakers are preparing regulatory action if the industry continues its current trajectory, potentially reshaping how quickly advanced AI systems can be deployed.

The details were first reported by Axios.