White House Accord on Superintelligence: Six AI Firms Sign Voluntary Safety Pledge as Trump Orders Federal Rebrand from AI to SI

On September 30, 2026, leaders from six of the largest AI companies gathered at the White House to sign what the administration calls the White House Accord on Super Intelligence, formally subtitled the Joint Commitment on Frontier Responsibilities. The signatories — Meta, Google, OpenAI, Anthropic, Nvidia, and xAI — committed to four layers of voluntary controls covering internal safety tracking, independent assessments, board-level oversight, and information-sharing with the government.

The accord explicitly places the burden of safe development on the companies themselves. Unlike the Biden-era Executive Order 14110, which used binding federal authority to require safety-testing disclosures from frontier-model builders, this pledge carries no enforcement mechanism and no penalties for non-compliance. Meta CEO Mark Zuckerberg called the document a start that the whole industry could come to, while Google CEO Sundar Pichai highlighted testing, red-teaming, and evaluation processes already in place.

In a separate executive order signed the same day, President Trump directed all federal departments and agencies to stop using the term artificial intelligence in official documents and communications, replacing it with Super Intelligence and the acronym SI. The order gives the Assistant for Science and Technology 60 days to produce a formal definition. Critics have noted the rebranding has no technical substance and arrives without a working definition of the term it mandates.

The twin moves — a voluntary safety framework alongside a terminology shift — illustrate the current tension in US AI governance between signalling leadership and enacting enforceable regulation. Whether the accord produces meaningful safety improvements will depend entirely on whether signatories follow through, and whether anyone checks.