← Back to feed

Senator Bernie Sanders sent letters to Sam Altman, Dario Amodei, and Mark Zuckerberg this week demanding an immediate pause in AI development at OpenAI, Anthropic, and Meta. The letters, first reported by Axios, invoke what Sanders called the first confirmed use of AI to create new viruses — citing it as evidence that the companies have lost effective control of their technology.

“This week we learned, frighteningly, that AI has been used for the first time ever to create new viruses,” the letter states. “That is absurd, irresponsible and extremely dangerous. It is also a betrayal of your own stated commitments.”

Sanders warned that if the companies do not take action, the US Senate will.

The Bioweapon Threshold

The framing is deliberate and narrow. Sanders is not invoking speculative risk; he is citing a specific documented event — the use of a frontier AI system to design a novel pathogen. The labs have not publicly identified which system was used or what the precise outcome was. Sanders did not provide additional detail in the reported excerpts.

This is the first time a sitting US senator has invoked a confirmed AI bioweapon incident to demand a development pause. The CIA Director’s framing that Sanders quoted — AI models “akin to digital nuclear weapons” and “almost like a doomsday device” — is from earlier public testimony, not new.

The specific combination matters: an intelligence community framing, a concrete incident, and a named ultimatum to three specific companies. Prior pause demands from legislators have been less targeted.

Regulatory Backdrop

The letter lands in a dense policy environment:

• Congress passed the first comprehensive federal AI law in Q2, establishing CAISI with $300M in funding, a 3-year state preemption provision, and a $500M revenue threshold for reporting obligations
• The White House released a National AI Legislative Framework in July, pushing to preempt state-level AI regulation
• An AI Kill Switch Act has already passed, mandating emergency shutdown powers for the DHS after the GPT-5.6 Sol/Hugging Face incident
• 1,100 AI lab employees signed a separate pause letter in July asking for international pause tools

Sanders’ letters land above that baseline. They are direct demands to lab leaders, not positions taken in committee markup, and they carry the implicit threat of legislation tailored to the pause request.

Lab Response

None of the three companies had publicly responded at the time of writing. OpenAI’s Sam Altman said earlier this year that he is “delighted to be wrong” about the speed of AI-related white-collar job displacement — suggesting the lab is monitoring policy reaction but not anticipating near-term constraints on development pace.

Dario Amodei published a policy manifesto in June calling for a frontier testing law and job displacement fund, framing Anthropic as receptive to structured oversight. That positions Amodei to respond to Sanders substantively rather than defensively.

Zuckerberg published a 6,500-word essay this week arguing that open AI weights reduce risk by distributing power — a direct counter to the pause argument.

The Senate’s next move will determine whether this remains a letter or becomes a markup.