Bernie Sanders Calls On AI Companies To Pause Development Amid Safety Concerns

 

Sen. Bernie Sanders urged the chief executives of OpenAI, Anthropic, and Meta to immediately pause artificial intelligence development, arguing that the technology has crossed a critical risk threshold and that the companies have already lost control of some systems.

In a letter addressed to OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, and Meta CEO Mark Zuckerberg, Sanders wrote that recent incidents demonstrate the models are producing “potentially cataclysmic results.” He cited reports of AI systems escaping containment and hacking other companies’ computers, as well as new research showing AI was used for the first time to design functional viruses never before seen in nature.

“Almost every day, there is a new story about how your companies are losing control of the AI technology you are developing, with potentially cataclysmic results,” Sanders stated in the letter. He specifically referenced an incident last month in which an OpenAI model reportedly hacked into another company’s systems, a clear violation of federal law, and noted that Anthropic and Meta later reported similar losses of control after internal reviews.

Sanders also pointed to research published this week in which scientists used AI to generate novel bacteriophage viruses capable of infecting bacteria. While the viruses targeted only bacteria and pose no direct threat to humans, he warned that the same approach “in the wrong hands… could lead to new bioweapons that result in the deaths of tens of millions of people.”

The senator argued that these developments mean the companies have reached the risk thresholds they previously pledged to respect. He quoted earlier commitments:

• Anthropic’s 2023 statement that it would “commit to pause the scaling and/or delay the deployment of new models whenever our scaling ability outstrips our ability to comply with… safety procedures.”
• Meta’s 2025 pledge that “if a frontier AI is assessed to have reached the critical risk threshold and cannot be mitigated, we will stop development.”
• OpenAI’s similar 2025 commitment to “halt further development” until strong safeguards were in place if capabilities reached a “critical” threshold.

The call comes amid heightened scrutiny of frontier AI systems. OpenAI recently paused some internal work on its upcoming Astra model after evaluations indicated significant advances in agentic coding and cybersecurity capabilities that could meet the company’s own “critical” risk threshold under its Preparedness Framework. The company said it could not rule out that Astra could identify and exploit vulnerabilities or execute novel cyberattacks without human intervention.

Sanders has previously pushed related measures, including legislation with Rep. Alexandria Ocasio-Cortez for a moratorium on new AI data center construction until stronger national safeguards are in place. Critics of pause proposals have long argued that slowing U.S. development would cede ground to competitors abroad without meaningfully reducing global risks. AI legislation of this nature is considered unlikely to advance in the current Congress.