Trump Signs Official AI Accord With Tech CEOs Favoring Self-Policing
President Donald Trump announced Sept. 29 that chief executives of the largest artificial intelligence companies had signed a voluntary White House accord favoring industry self-policing over new federal rules.
At a lunch with House Speaker Mike Johnson, attendees including Anthropic chief executive Dario Amodei and other industry leaders discussed AI oversight and the data-center buildout powering large models.
The president called the document a “constitution” for the industry, though the White House describes it as an “accord,” according to Forbes.
The 308-word text commits signatories to internal and external reviews of AI systems, according to AP, while he said he would never “stifle” the industry with regulation, framing self-policing as the preferred path despite recent AI-safety alarms from researchers and lawmakers.
What The Accord Actually Says
The document is short enough that critics questioned whether it carries binding weight beyond a press-friendly signing ceremony. It commits companies to reviews but sets no enforceable standards, timelines or penalties for noncompliance, according to Reuters. This was not a product release, it identifies no user-facing availability, pricing, rate limits, regions or tiers.
Trump also backed continued expansion of data centers, the power-hungry facilities that train and run large AI models, pairing infrastructure growth with voluntary safety pledges rather than mandatory limits. The signing demonstrates a commitment, not an established change to how a company’s AI product operates on an ordinary day.
He had recently dismissed warnings about unregulated AI as a “hoax,” according to USA Today, making Tuesday’s accord a shift toward acknowledging risk while rejecting government-imposed rules. Calls for a regulatory slowdown had reached what CNBC described as a fever pitch among lawmakers and researchers, concerned with AI risks and possible limitations. Whether an unenforceable, industry-drafted pledge changes actual company behavior remains untested, and the accord has no independent verification mechanism.
Read Next: Short Anthropic ETF Filing Seeks Double Inverse Daily Return
