Sanders and Bannon share a Washington stage to demand AI limits

Senator Bernie Sanders and Steve Bannon will speak on Tuesday at the same Washington event calling for limits on advanced artificial intelligence. The Future of Life Institute titled the gathering the Pro-Human Assembly. The AFL-CIO president, Liz Shuler, and Walter Kim, president of the National Association of Evangelicals, are also on the programme. Other listed speakers include Sen. Marsha Blackburn, Reps. Greg Casar, Chip Roy, Lori Trahan and Don Beyer, and New York Assemblyman Alex Bores.

Sanders and Bannon are not sharing a joint statement and are not appearing as a ticket. They are using the same room because both now treat frontier AI labs as a political problem rather than a sector to cheer. Sanders wrote to the chief executives of OpenAI, Anthropic and Meta in August and told them to pause work on systems humans cannot control. He and Casar have called for a ban on superintelligence until a new federal agency exists to police it. He has cited model-enabled hacking and virus-design papers as evidence that the threshold he feared has already been crossed.

Bannon framed the same fight in populist language on his Monday podcast. He said Congress should refuse any bill that stops states from writing their own AI rules. Reuters has reported that senators are debating exactly that kind of pre-emption. Bannon said Americans will not be “supplicants to these tech oligarchs.” The overlap is the word human. The Institute agenda lists keeping humans in charge, avoiding concentration of power, protecting human experience, and assigning responsibility for AI systems.

The midterm calendar is the reason the room exists this week. AI safety has moved from laboratory letters into campaign copy. Sanders treats the issue as a labour and existential risk. Bannon treats it as a fight with coastal firms and with any federal statute that would lock California or Texas out of the field. Labour and evangelical groups in the same hall give both men a constituency that is not the usual tech-policy circuit.

None of this settles the legislative text. A pause on training runs, a ban on systems above a stated capability, a new agency, and a shield for state law are four different bills. The companies Sanders named have not agreed to stop. The White House has called extinction talk a hoax in other settings this month. France is separately pushing an AI session at the UN Security Council during the General Assembly.

What Tuesday adds is a public record that the loudest socialist in the Senate and one of the loudest voices on the MAGA right now use the same verb about the same labs: stop, or at least slow, the next training run. Whether that alignment lasts past the midterms will depend on the pre-emption bill and on whether either man can point to a vote, not only a speech.

The Institute will publish the remarks. The test for readers is whether any of the speakers name a number, a model class, or an enforcement body that did not already exist in their prior letters. If they do, the event is a policy marker. If they do not, it is a date on which two enemies agreed the industry has outrun the law.