Trump backs voluntary AI self-regulation after White House tech summit

President Donald Trump met leading AI executives at the White House for a private lunch and discussion on regulation. Nearly 30 technology leaders attended, including bosses from Nvidia, Meta, OpenAI, Google, Anthropic and Elon Musk of xAI. The group discussed guardrails for artificial intelligence, which Trump now calls superintelligence, or SI.

After the private meeting, the executives briefly spoke to reporters. Trump then appeared unexpectedly and took questions. He said the group had reached an agreement and described it as "almost like a constitution", adding that the biggest people in the world signed it and he signed it as president. He called it a form of protection.

The White House Accord on Super Intelligence was posted on Truth Social. It is a single page. Six companies agreed to four main things:

• implement internal controls so AI models do not hack systems in unintended ways;
• create internal watchdog teams to monitor and detect issues;
• partner with an external auditor to check controls;
• install independent committees on their boards to receive reports from those teams.

The agreement is voluntary and morally binding, not law. It says that over time it may make sense to codify the steps into laws or regulations.

Trump ruled out a joint venture with China to control AI growth, saying it would give away tech secrets to America's main economic rival. He said AI companies would police themselves. "There's going to be a tremendous self-policing aspect," he said. He also said his administration might form a committee of about 10 people from the group to watch over the sector. He repeated that the Department of Justice and the FBI provide some regulation.

The meeting came after safety warnings. Anthropic warned would-be investors that its business has the potential to wipe out humanity. OpenAI delayed the release of its latest AI model because of safety fears. Its chief executive used the summit to unveil a new Codex security cloud. At the White House, Anthropic's boss said the technology has very real risks and the mechanism to address them is still under discussion. Another executive said the accord was a start and that the industry needed to work together. Elon Musk thanked Trump for the SI rebrand and said the most likely outcome for AI is incredibly beneficial.

Press questions outside the meeting went unanswered by some executives. Reporters asked whether self-regulation was dangerous and whether the technology was safe. One executive said he would love to talk but could not comment after such a meeting. Another joked about trying to self-regulate his own addictions. Critics said the voluntary approach leaves tech billionaires marking their own homework. One analyst described the deal as a wartime posture in the race with China, saying a kill switch, security by design and impact reviews would be needed to make it work.