A White House meeting on AI regulation ended with tech leaders agreeing on an AI "constitution." Stuart Russell, president of the International Association for Safe and Ethical AI, joins "The Daily Report" to break it down.
Page A6From§Eachthe late evening edition — 30 September 2026
Trump's AI pledge is already 'entirely unenforceable,' and the industry's own money just walked out on it.
As it ran on the front
"Follow the money" is supposed to be a metaphor. Try it literal.
That is the model: a super PAC moves the cash fast, with a media buyer in between to keep a donor's fingerprints off the ad. Ten days later, on the 15th, New York state assemblyman Alex Bores went on CBS to lay out what AI safety measures could look like as the regulation debate heats up. A week after that, on the 22nd, a UN-backed scientific panel told world leaders the safeguards now in place are "unraveling." Secretary-General António Guterres welcomed the finding and, separately, asked outside experts — including ones from the labs themselves — to help check the work. None of it answered the question a Senate hearing had already put on the record: that a single AI glitch in military command systems could trigger a nuclear exchange. Congress still has no AI safety law on the books. Not one agency has been assigned to catch that glitch before it becomes the exchange.…
…(cont) Then came September 30th, and the industry had its answer ready: it would police itself. Trump took a pledge from the AI companies that they would self-regulate. House Minority Leader Hakeem Jeffries called the deal "entirely unenforceable" that same day, telling reporters that letting "the industry writ large" police itself "is not the right" approach. He had better timing than he knew. That same day, Greg Brockman, OpenAI's president, told his own company that Leading the Future, the AI industry's super PAC, had become "a distraction," and pulled his second $25 million out of it. The pledge to govern the industry and the fund built to keep Congress from governing it collapsed in the same news cycle, from the same side of the table.
By October 1st the pattern had a second witness. The New York Times noted Trump's voluntary policing plan runs the same script Biden tried, recycled while the companies keep failing to put safeguards on the technology they already sell. Stuart Russell, who leads the International Association for Safe and Ethical AI, went on CBS and said, plainly, he does not buy it.
Nobody passed a law. Somebody signed a pledge. The money built to protect that pledge from a law walked out the door the same week the pledge was signed.
“They sell self-policing like it's a new product, but watch the receipts: the same week the president gets his pledge, the guy who helped build the industry's super PAC pulls his second $25 million out of it. That's not a distraction, that's a bill paid in full — once you've got the deal, you stop renting the insurance. Jeffries called it unenforceable; the only enforcement anybody ever built here was the money, and now that's walking too.”
“Voluntary agreements work, that's just how the market self-corrects, OK — the pledge is morally binding, which is a real category, I've seen the memo. Did I say morally binding, that doesn't sound like me. Look, self-regulation moves faster than Congress ever could, and if a company pulls $25 million out of a PAC, that just means the system's already working — nobody needs to buy influence once the deal's done.”
Earlier in this story
- 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
- Senate Democrats kill data center bill as too weak, their candidate cashes its industry's checks. By Lou · the matinee edition, 30 September 2026
- Senate loses war powers vote, then blocks a vote on the ad money six days later. By Ruth · the noon edition, 30 September 2026
- Democrats finally fund a five-point Texas lead; a friendly outlet brands him a liar the next day. By Sal · the breakfast edition, 30 September 2026
- Only the senator whose own constituent was killed votes against investigating the killing By Sal · the sunrise edition, 30 September 2026
The receipts
President Trump’s plan to let A.I. companies police themselves comes amid mounting examples that they are failing to put safeguards on the technology.
House Minority Leader Hakeem Jeffries (D-NY) dismissed the agreement President Donald Trump reached with top executives of artificial intelligence as “entirely unenforceable.” Jeffries, speaking to reporters on Wednesday, said the “notion of entering into a voluntary agreement that is entirely unenforceable and that allows the industry…
Greg Brockman, OpenAI’s president and co-founder, said internally that the super PAC, Leading the Future, had become a “distraction” for the A.I. company.
· UN panel calls for guardrails on AI as current safeguards are 'unraveling' from the morgue, 22 Sep 2026
· AI expert on the path to tech safety and security from the morgue, 15 Sep 2026
· Trump's MAGA Inc. super-PAC funnels $10 million into Texas Senate race from the morgue, 5 Sep 2026
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.