Tech executives sign voluntary AI safety accord with Trump, who orders agencies to call AI 'SI'

Six tech executives signing an AI safety accord with President Trump will police their own companies under a voluntary framework announced Tuesday, with no binding government enforcement built in. The White House Accord on Superintelligence asks each signer to build internal controls, an internal review team, an outside auditor and a board committee that reviews the auditors' findings. The accord targets risks in cybersecurity, biosecurity and chemical threats, where the companies pledged to monitor whether their models behave as intended and fix problems quickly. Signers included Sundar Pichai, Dario Amodei, Mark Zuckerberg, Greg Brockman, Elon Musk and Jensen Huang.

Trump also signed an executive order renaming the technology "Super Intelligence," or SI, across federal agencies. "We've changed the name, Steve, officially to SI," he told reporters.

A reporter asked whether the accord was binding; the answer was "morally binding." One executive said the mechanism for addressing the risks is still under discussion and that the industry needs to "work together to make sure that we can win and we can win safely."

Trump said he'll name a superintelligence czar (a lead official): "I'll be naming, I'd say, over the next 3 or 4 days."

No one outside the event is quoted.