Six major AI companies have signed a non-binding White House agreement on safety standards, while a new executive order mandates a change in government terminology for the technology.

President Donald Trump has secured commitments from six leading AI companies to follow a set of non-binding safety guidelines titled the "White House Accord on Superintelligence." Meta, Google, OpenAI, Anthropic, Nvidia, and XAI signed the document, which outlines voluntary internal controls and audit processes for frontier model development. The agreement does not mandate specific enforcement actions or provide strict definitions for its safety requirements.

Accord Requirements and Oversight

The accord, subtitled the "Joint Commitment on Frontier Responsibilities," suggests that companies implement four layers of controls, including internal monitoring for cybersecurity and biosecurity risks. It also calls for the empowerment of internal oversight teams, the engagement of independent external auditors, and the designation of a board-level committee to manage compliance. The document does not define what constitutes "robust" controls, nor does it specify the frequency of required audits or evaluations. Participating companies have agreed to meet regularly to establish their own safety standards and best practices, leaving the possibility of future codification into law open.

Government Terminology Mandate

In a separate action, President Trump signed an executive order requiring federal agencies to replace the terms "Artificial Intelligence" and "AI" with "Super Intelligence" and "SI" in all official communications and documents. The order tasks the Assistant for Science and Technology with developing a formal definition for "Super Intelligence" within 60 days.