On September 29, 2026, President Donald Trump hosted a lunch at the White House with CEOs of six major AI firms. After the meeting the companies posted a two‑page document on Trump’s Truth Social account. The paper, titled “White House Accord on Superintelligence: Joint Commitment on Frontier Responsibilities,” lists voluntary safety steps the firms will take for their most advanced models.
What the accord says
The accord asks each signatory to put in place four layers of oversight: robust internal monitoring of model behavior, a dedicated internal team to run the controls, regular independent audits, and a board‑level committee that reviews audit reports. Companies also agree to meet regularly to share standards and best practices. The language is vague – it does not define how strong the controls must be, how often audits should happen, or what “robust” means.
Who signed and what else happened
The six signatories are Meta (Mark Zuckerberg), Google (Sundar Pichai), OpenAI (Greg Brockman), Anthropic (Dario Amodei), Nvidia (Jensen Huang) and XAI (Elon Musk). In the same day’s press event, Trump signed an executive order that forces every U.S. federal agency to replace the term “artificial intelligence” with “Super Intelligence” (SI) in all official communications. The order does not provide a definition for SI and gives the White House’s science office 60 days to work out the details.
Why it matters
For everyday users, the pledge means the biggest AI makers say they will watch their own systems more closely, but the rules are voluntary and vague. The new government label won’t change how the technology works, but it may shape public perception and future policy discussions.
