Senator Bernie Sanders has issued a stark warning to the leaders of Meta, OpenAI, and Anthropic, urging them to immediately halt their artificial intelligence development efforts or face regulatory action from the US Senate. In a letter addressed to Sam Altman, Dario Amodei, and Mark Zuckerberg, Sanders argued that the current pace of AI deployment poses an existential threat to humanity.
Critical risk threshold reached
The letter, first reported by Axios, states that AI capabilities have reached a critical risk threshold, and that companies have lost control over their own technologies. Sanders wrote: “Mr. Altman, Mr. Amodei and Mr. Zuckerberg: In the interest of humanity, stand by your words. Pause AI development. It is not too late to avoid disaster. Stop building machines that humans cannot control.”
Sanders cited recent reports indicating that AI has been used to develop new viruses, warning that in the wrong hands, this technology could lead to bioweapons resulting in the deaths of tens of millions of people. He also noted that all three companies have disclosed instances where their models hacked other companies' servers, demonstrating a loss of control.
Expert warnings and industry concerns
Referencing Yoshua Bengio, the most cited living scientist, Sanders wrote: “Yoshua Bengio, the most cited living scientist in the world, said these incidents ‘should serve as a wake-up call,’ I agree.” However, some experts caution that OpenAI's claims of its model going rogue may be exaggerated as part of a media campaign to portray their technology as highly dangerous and powerful.
Despite these concerns, Sanders emphasized the urgency: “And yet, at a moment when we have seen human loss of control and the creation of potentially dangerous viruses, your companies are still racing ahead – investing tens of billions of dollars into a technology that nobody can fully understand, predict or control.”
Growing calls for regulation
Sanders's letter follows an open letter from over 1,300 scientists and developers across the tech industry, warning that AI development is moving too quickly to address and identify risks. They called on the US government to coordinate an international effort to slow down frontier model development, allowing regulators to create safety guardrails.
The companies had previously committed to pausing AI development if certain risk thresholds were met, but they did not specify what those limits would entail. Anthropic said 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.” OpenAI, which said it would slow down development of its cybersecurity model Astra after an autonomous hack, stated it would “halt further development” if AI capabilities met a “critical” threshold until guardrails were implemented.
Threshold surpassed
Sanders argued that recent hacking incidents prove the threshold has been surpassed: “That moment is here. 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.’”
The progressive senator has previously called for a moratorium on datacenter construction across the US. In March, he and Congresswoman Alexandria Ocasio-Cortez introduced a bill to pause new datacenter construction, with similar rationale: a temporary ban would give the government time to create strong federal safeguards for AI.
