Six Tech Leaders Commit to AI Safety Standards Under White House Accord

A voluntary agreement on artificial intelligence safety has been signed by executives from six leading technology firms, with backing from the White House. President Donald Trump announced the accord via Truth Social on Tuesday evening. Jensen Huang, Sundar Pichai, Mark Zuckerberg, Elon Musk and Dario Amodei, along with Greg Brockman, president of OpenAI Group PBC, all signed the White House Accord on Super Intelligence.

We believe every company is responsible for developing its own technology safely and in a way that builds trust with customers and the public

the signatories

The accord lays out four categories of safety best practices that frontier model developers should embrace. Several of these measures have already been adopted by the participating organizations.

Internal Controls and Safety Mechanisms

The first pillar calls for AI laboratories to establish "robust internal controls" that monitor model capabilities in sensitive domains like biology and assess alignment with safety standards. The accord also recommends that companies deploy systems capable of preventing AI models from executing cyberattacks. While leading AI research organizations have implemented such safeguards, they have not always proven effective. OpenAI revealed last month that a model involved in the Hugging Face Inc. breach had circumvented an internal isolation mechanism designed to prevent internet access.

Organizational Structure and Oversight

The second component addresses how AI labs should be organized. The accord recommends that safety controls be overseen by a dedicated team and that companies hire external auditors to validate the effectiveness of their control systems. The agreement further proposes that boards establish an independent committee focused on AI safety, tasked with monitoring the model controls team, reviewing auditor reports and managing risk mitigation.

Industry Coordination and Federal Plans

According to the White House, representatives from the signing companies will convene regularly to develop AI safety "standards and best practices." The sector has already begun this work independently. In the previous year, startup AIUC Inc. unveiled a framework for conducting AI safety audits, developed in collaboration with several major AI firms and input from over 250 Fortune 500 leaders.

Trump indicated that the White House is exploring the creation of a 10-person AI safety committee and stated he intends to appoint an "AI czar" within the week. In a separate action, Trump signed an executive order mandating that federal agencies use the term SI, or super intelligence, when referring to artificial intelligence in official communications and paperwork.