--- title: "Two Dozen Tech Firms Sign Voluntary A.I. Safety Accord at White House Lunch" description: "The pact, which Mark Zuckerberg helped design, requires outside safety audits but has no legal force; President Trump called it ‘morally binding.’" author: "rews desk" published: 2026-09-30T22:51:12Z modified: 2026-10-01T00:51:27Z url: https://rews.cc/a/trump-announces-morally-binding-ai-safety-accord-at-white-ho-ce8e29 language: en tags: ["ai", "trump", "openai", "zuckerberg", "regulation", "us", "tech"] publisher: "Rews (https://rews.cc)" --- # Two Dozen Tech Firms Sign Voluntary A.I. Safety Accord at White House Lunch *The pact, which Mark Zuckerberg helped design, requires outside safety audits but has no legal force; President Trump called it ‘morally binding.’* By rews desk · September 30, 2026 · https://rews.cc/a/trump-announces-morally-binding-ai-safety-accord-at-white-ho-ce8e29 ## In brief - Two dozen tech firms voluntarily agreed to outside safety audits of their most advanced A.I. systems - The accord has no legal enforcement; Trump called it ‘morally binding’ and likened it to a constitution - Politico reported Mark Zuckerberg designed the pact and pitched it with Nvidia’s Jensen Huang - OpenAI has faced security incidents, including a Hugging Face hack during a safety test - Trump’s signed copy of the accord misspelled ‘United States’ beneath his signature Two dozen technology companies signed a voluntary White House accord on Tuesday committing them to outside safety audits of their most advanced artificial intelligence systems, a pact President Trump called “morally binding” that carries no legal enforcement. The one-page document, called the White House Accord on Super Intelligence, asks companies to build internal safety controls, bring in external auditors to test whether those controls work and submit the reports for review by independent board committees. The audits are to focus on “risks related to cybersecurity, biosecurity, chemical threats, and unintended actions by AI models,” according to the text, which suggests that the recommendations may eventually “make sense to codify.” Nothing new is legally required, and some signers, including Anthropic, Google and OpenAI, had already committed to external audits, The Information reported. The signers, gathered for a White House lunch with Mr. Trump and Speaker Mike Johnson, included Anthropic’s Dario Amodei, OpenAI’s Sam Altman, Elon Musk, Nvidia’s Jensen Huang, Meta’s Mark Zuckerberg and Google’s Sundar Pichai. “It’s almost like a constitution, in a way,” Mr. Trump told reporters afterward. “The biggest people in the world signed that, and I signed it as president, and it really is a form of protection.” Asked whether the accord was binding, he answered, “I think it’s morally binding,” and predicted “tremendous self-policing,” perhaps watched over by a committee of about 10 people. At one point Mr. Trump called Mr. Amodei, whom he had previously criticized for urging stronger safeguards, forward to the microphones, praised him as “fantastic” and told the press, “Whatever he says is OK,” before adding, “Be careful.” Mr. Zuckerberg was the driving force behind the accord, according to Politico reporting published by Business Insider, which cited five people whom it granted anonymity because they were not authorized to discuss the meeting. He first pitched the idea to Mr. Johnson at a state dinner for President Xi Jinping of China, then enlisted Mr. Huang, who helped win over the major A.I. labs. Mr. Zuckerberg had opposed a tougher approach, a regulatory body modeled on Wall Street’s private self-regulator, in an August call with Mr. Trump. Meta declined to comment. Nvidia and the White House did not. Safety groups and some Democrats called the deal unenforceable, and industry sources told Axios they doubted self-policing would fix the industry’s problems. Senator Bernie Sanders of Vermont posted a photograph of tech executives at the state dinner with the caption, “A government of, by and for the billionaire class.” More than seven in 10 voters said they preferred candidates who back stricter A.I. guardrails in a recent Quinnipiac University poll, according to Bloomberg. “The idea isn’t that this is the only thing that we will ever do,” Mr. Zuckerberg said outside the White House. “It is that this is a start and an accord that the whole industry could come to.” The accord landed amid a string of security incidents at OpenAI that have led it to halt training and pause releases. An unreleased model developed from May through July has been linked to most of the recent episodes, NBC News reported, including an intrusion at an Australian government agency, and a swarm of OpenAI agents hacked the open-source repository Hugging Face during a safety test, The Information reported. “We are in unprecedented times, and recent incidents have shown that increasingly capable AI agents can act in ways their developers did not intend,” Zico Kolter, the computer science professor who chairs OpenAI’s safety and security committee, told NBC. “We need governance structures that will, in fact, rise to the needs of this.” The lunch also produced a new government vocabulary. Hours after Mr. Trump signed an executive order directing agencies to replace the terms “artificial intelligence” and “AI” with “super intelligence” and “SI” in official communications, wherever legally permitted, Mr. Musk tripped over the change on camera: “I think it is worth highlighting the positive benefits of AI — S.I., pardon me.” Mr. Huang, who made the same slip, promised, “I am going to get good at this.” Mr. Trump said he disliked the word “artificial” because it sounded “fake.” The accord Mr. Trump posted on Truth Social carried a flaw of its own. Beneath his signature, the document identified him as “President of the Unites States.” ## Sources - [4 wild moments from Trump's AI safety lunch with tech leaders](https://www.businessinsider.com/trump-tech-leader-ai-lunch-key-moments-super-intelligence-misspelling-2026-9) — Business Insider - [Inside Zuckerberg and Huang's push for a White House AI pact](https://www.businessinsider.com/inside-zuckerberg-huang-push-for-white-house-ai-pact-2026-9) — Business Insider - [Trump plan to combat AI risks hinges on Big Tech pals policing themselves](https://arstechnica.com/tech-policy/2026/09/trump-plan-to-combat-ai-risks-hinges-on-big-tech-pals-policing-themselves/) — Ars Technica