Trump Lets AI Companies Police Themselves: What It Means
On September 29, 2026, President Donald Trump signed a new executive order on AI and announced a voluntary promise by six tech chiefs to watch over their own most powerful AI systems.
The order is titled 'Inaugurating the Era of Super Intelligence' The White House. A related paper is titled 'Fact Sheet: President Donald J. Trump Inaugurates the Era of Super Intelligence'. Both are dated September 29 and listed in the White House briefing room under executive orders and fact sheets.
The industry paper is the Joint Commitment on Frontier Responsibilities. The signers are Google's Sundar Pichai, Anthropic's Dario Amodei, Meta's Mark Zuckerberg, OpenAI's Greg Brockman, xAI's Elon Musk and Nvidia's Jensen Huang The Verge. All six gave press comments at Trump's request praising the self-policing plan.
A voluntary accord
The promise is described as "morally binding." That means companies agree to set their own common-sense safety rules. Trump said on September 29 that tech chiefs had agreed to set voluntary standards for AI Reuters, and that the industry can police itself after signing the accord.
His wording was direct. "I think I'm seeing tremendous self-policing. And they understand that they have to self-police" NPR. He spoke about AI companies self-policing to reporters outside the West Wing PBS.
No new federal rule followed. Trump answered concerns about lack of government enforcement by saying U.S. companies are "dominating China." He said "the smartest people in the world" are "watching over each other" instead of bringing in federal AI rules.
Trump hosted a meal with Big Tech leaders on Tuesday before talking about his AI announcements. He held a meeting with tech chiefs on AI on September 29, 2026 Roll Call. That day he also said top tech chiefs had signed a new promise for better self-policing of their companies' AI CNN.
Renaming AI
Trump also spoke about words. He said he disliked the word "artificial" in AI because it sounds "fake" Business Insider. He separately said the word "artificial" isn't helpful NPR.
Trump had earlier called AI safety fears a "scam" CNBC. He said his earlier claim that AI safety fears were a "hoax" cooked up by Democrats no longer applies after renaming the technology to "Super Intelligence."
The administration placed the order in a longer chain. In March 2026, the White House shared a National Super Intelligence Policy Framework, as described in the September fact sheet The White House.
What voluntarism means in practice
For plain context, this plan covers the full AI pipeline. That includes early training, added training to shape behavior, and safety checks when the system answers users. Voluntary promises can move fast because they skip long rule-making. Like drivers agreeing to speed limits with no police on the road, they depend on each company to follow through.
The broader context here is that fast does not always mean consistent. In my view, the test is not whether labs do safety tests. They do. The test is whether common-sense rules turn into shared practice that can be compared across companies.
In my view, the clear signs will be simple. Which safety tests are made public. Which model weights, the core settings, are held back. Which smart agent features need extra review. And how labs share warning signs when something goes wrong.
Watching my own children grow up with PCs, the web, phones and cloud services left me hopeful about how people adapt, and careful about gaps in oversight. People tend to adopt faster than rules are written.
What this could enable, if companies follow through, is more U.S. AI building under light federal rules. The administration is counting on peer pressure among lab leaders, plus demands from business buyers for checks and clear origins, to stand in for formal rules while keeping pace with Chinese rivals.
