{beacon} Technology Technology The Big Story Debate over AI regulation pushed past midterms The debate over how to regulate artificial intelligence is being pushed past the November midterms, much to the dismay of some lawmakers. © Mariam Zuhaib, Associated Press Our colleague Sudiksha Kochi's latest Insider piece breaks down the flurry…
Page A3From§Eachthe coffee break edition — 1 October 2026
Florida sues to force OpenAI to do what Trump's revived self-policing pledge already promised.
As it ran on the front
On September 21st, a defense writer used a Washington Examiner column to describe an AI glitch that had nearly started a war, arguing that nuclear command and control could not survive being treated like a feature still in beta. The next day, a UN-backed panel told the world's governments the safeguards that already exist are "unraveling." None of that moved the people positioned to write an actual rule: on September 28th, The Hill reported the debate over AI regulation in Congress had been pushed past the midterms — past this term, meaning the industry keeps writing its own rules on its own clock for a while longer, not the public's.
Two days after that the administration picked up where the warnings left off. On September 30th the New York Times reported Trump revived the same voluntary AI safety pledge the Biden administration had already tried — same companies, same ask, same missing referee.…
…(cont) By Thursday the pledge had company it did not ask for. RTE reported the federal government's own chatbot, built to answer citizens' questions, had been contradicting Trump's claims until its answers were edited — corrected until the record matched the man, not what the chatbot had found on its own. That is the administration's software, inside the administration's own building, and it still needed a hand on the delete key. Late night noticed the same week, the Times reported, and so did a computer science professor CBS asked directly whether the pledge to self-police would hold. A second CBS interview asked an AI expert what a rogue model can trigger without a human override — the doomsday scenarios, discussed on camera as a live category, not a hypothetical for a conference panel.
Thursday afternoon the test case walked in. CBS reported Florida's attorney general is trying to stop OpenAI from developing its products until a court rules them safe — a state reaching for the authority the federal pledge was supposed to make unnecessary. University of Florida law professor Derek Bambauer told CBS the answer to whether a state can regulate AI is yes. That is the entire argument the self-policing pledge exists to preempt, now running in a Florida courtroom instead of a congressional committee, because the committee was told to come back after the midterms.
“Every time the companies say "trust us," somebody with a law degree or a defense contract has to go check, and the checking keeps landing in a Florida courtroom instead of a Senate hearing room, because the hearing room already told everybody to come back after the midterms. Watch the hands: the pledge gets the headline, the editing of the government's own chatbot gets buried in a tech wire story, and the guy asking whether a rogue model can reach the nuclear command system gets a video segment instead of a subpoena. That's not an oversight, that's the order of operations.”
“Look, a voluntary pledge means the industry takes safety seriously without the dead hand of fifty different state regulators — Florida going after one company sets exactly the kind of patchwork that chills innovation. And the chatbot thing, that's just quality control, every product gets edited before — wait, I'm defending the editing of government answers to match the president now? That's not usually my lane either.”
Earlier in this story
- Trump revives Biden's AI self-policing pledge days after a Senate warning on nuclear command and control. By Sal · the milkman edition, 1 October 2026
- Trump's AI chatbot disputes his claims, then stops disputing them. By Sal · the wee small hours edition, 1 October 2026
- AI industry polices itself again, same plan Biden tried, same week Senate hears it could trigger nuclear war. By Sal · the midnight edition, 1 October 2026
- Trump's AI pledge is already 'entirely unenforceable,' and the industry's own money just walked out on it. By Sal · the late evening edition, 30 September 2026
- Congress passed no AI law, so no senator could say who stops a nuclear AI glitch. By Sal · the early evening edition, 30 September 2026
- AI CEOs sign 'morally binding' pledge; FTC opens investigation into signers days later. By Sal · the supper edition, 30 September 2026
The receipts
Florida's Republican attorney general wants to stop OpenAI from developing its artificial intelligence products until a court determines they are safe. University of Florida law professor Derek Bambauer weighs in.
This page is a back-issue: the story as it ran, receipts as they were. The current edition is at the front. The byline is a pen name for a column drafted by a machine and checked by the editor: how this is made.