Donald Trump says leaders of major artificial-intelligence companies have signed a voluntary accord that will include internal and external reviews of the technology. The text of the agreement has not been released, leaving its legal force, scope and enforcement mechanisms unclear.

According to the Associated Press, Trump said roughly 10 people would be named to a committee to watch over the effort and described the accord as morally binding. The announcement followed a White House meeting with technology executives.

Google chief executive Sundar Pichai described the arrangement as a set of processes and controls comparable to corporate financial controls. Meta chief executive Mark Zuckerberg said it would involve internal risk reviews, external auditors and evaluators, and board-level review of reports.

Those descriptions point to a governance framework rather than a new law. Voluntary commitments can establish useful common practices, but the absence of a published text makes it impossible for outside researchers, users or lawmakers to assess which systems are covered, who chooses the evaluators or what happens when a company fails a review.

Anthropic chief executive Dario Amodei said the technology carried real risks and that the mechanism for addressing them was still under discussion. That caveat is central: the White House announcement may be a first step toward a shared safety process, but the public evidence so far is an announcement, not a complete accountability regime.