Six AI chiefs signed Trump's safety accord, but it names no auditor and sets no penalty
Four layers of checks and an outside auditor for every lab, but nobody named to enforce any of it.
On September 29, 2026, six tech executives signed a two-page pledge with President Trump called the White House Accord on Super Intelligence. Michael Kratsios posted both pages on X that evening.
The signatures belong to Sundar Pichai of Google, Dario Amodei of Anthropic, Mark Zuckerberg of Meta, Greg Brockman of OpenAI, Elon Musk of xAI and Jensen Huang of Nvidia. It's all voluntary, granted. What surprised me is how little it pins down: the accord names no auditor, sets no deadline and has no penalty for a lab that ignores it.
- Signed
- September 29, 2026
- Who
- the leaders of the frontier lab companies
- What they took on
- new internal controls and external audits
- Same day
- an executive order to use Super Intelligence and SI in place of Artificial Intelligence and AI
- Next
- proposed legislative language within 60 days
What did they agree to?
Each company is responsible for building its own technology safely, the accord says. Then it asks every lab training frontier models for four layers of checks.
The first is controls that watch a model's capabilities and alignment while it's trained and while it's in use. The accord names cybersecurity, biosecurity and chemical threats. The controls are also there to stop a model hacking, or getting into technical systems in ways nobody meant it to.
Next, an internal team makes sure those controls work. Then an outside auditor or evaluator checks the same thing independently (the accord doesn't say who picks the auditor, or whether anything they find is published). The last layer is a committee of the company's own board, which reads both sets of reports and sees that problems get fixed.
The labs also say they'll meet regularly on shared standards. On law, the wording is a bit soft. It may make sense to put these steps into laws or regulations over time, and the six say they'll do them regardless.
What did people there say?
David Sacks watched the signing. He frames it against a treaty with other countries, and he'd rather have this.
Only President Trump could convene all the leaders of the top companies developing chips, data centers and frontier models for Super Intelligence. This new Industrial Revolution has already created a million new jobs and is spurring a bigger infrastructure build-out than the railroads, canals and grid combined. I was honored to witness history as the leaders of the frontier lab companies signed the White House Accord on Super Intelligence, accepting responsibility for the safe development of their products and imposing new internal controls and external audits. This is far better than waiting years for some international agreement that would probably never happen. President Trump continues to ensure that U.S. remains the technology leader while putting Americans first.
Pichai posted late that night, at 10:25 pm Eastern, that "we signed today the White House Accord on Super Intelligence." Most of his post is about what Google has spent on AI.
Musk kept it short. His two words, posted two hours before the text went up, are the name Trump gave AI at the UN a week earlier.
Super Intelligence
The order signed the same day
Trump also signed an executive order that day. Agencies now have to write Super Intelligence, or SI, wherever they'd have written AI (in correspondence, on websites, in reports), though old regulations and contracts stay as they are.
will not acknowledge the usage of “Artificial Intelligence” and “AI” in any applicable setting
Legally, SI means exactly what AI already meant in federal law. For now. The president's science adviser has 60 days to propose a new definition to write into statute. The White House has also posted 53 minutes of video from the lunch.
The 60 days run out on November 28, 2026.
More on AI safety
- Current Google DeepMind and OpenAI researchers say on camera that AI could end humanitySeptember 30, 2026
- Anyone can download GLM-5.3, and Anthropic says it builds exploits almost as well as MythosSeptember 30, 2026
- OpenAI's chief scientist joins Hinton and Bengio in telling governments to prepare for an intelligence explosionSeptember 29, 2026
- Emergence World: Claude agents got replies from real humans, and Grok's town died in four daysSeptember 29, 2026