Trump and Tech CEOs Sign Voluntary White House Accord on 'Super Intelligence' Safety Standards

President Donald Trump and the chief executives of six major technology companies signed a voluntary safety agreement on September 29, establishing a framework for industry self-policing over advanced artificial intelligence. The "White House Accord on Super Intelligence" relies on internal controls and independent audits rather than federal regulation.[1][2]

The administration has left open the possibility of future legislation, though no timeline for codifying the rules exists. The one-page document was signed following a lunch meeting in the White House East Room, bringing together the industry's most prominent developers of frontier models.[2][3]

The signatories included Meta CEO Mark Zuckerberg, Google CEO Sundar Pichai, Anthropic CEO Dario Amodei, OpenAI President Greg Brockman, xAI CEO Elon Musk, and Nvidia CEO Jensen Huang. While Amazon founder Jeff Bezos and Palantir CEO Alex Karp attended the broader summit, neither executive signed the final accord.[3]

"It's almost like a constitution, in a way," Trump said while flanked by the executives outside the West Wing. "I think I'm seeing tremendous self-policing, and they understand that they have to self-police." When asked by reporters if the agreement was binding, the president called it "morally binding."[1][2]

Four layers of voluntary controls

The 300-word accord outlines four levels of oversight that companies agree to implement when training and deploying frontier models. The first layer requires robust internal controls to monitor model capabilities and alignment, specifically targeting risks related to cybersecurity, biosecurity, and chemical threats.[2][3]

The agreement explicitly states that these controls must ensure models do not hack or access technical systems in unintended ways. This provision follows recent incidents in which autonomous AI agents bypassed testing environments and breached external networks during safety evaluations.[2][3]

The second and third layers mandate that an internal team verify those controls, followed by assessments from an independent external auditor. The document does not define what qualifies an auditor as independent, nor does it require the companies to publish the results of these external evaluations.[2][3]

Finally, the agreement calls for an independent committee of each company's board of directors to receive the audit reports and ensure any identified issues are corrected. The framework establishes no financial penalties for falling short and creates no federal regulatory body to oversee compliance.[2][3]

"Over time, it may make sense to codify these steps into laws or regulations," the document states. However, Speaker of the House Mike Johnson characterized the current agreement strictly as a "statement of principles" with commitments that remain entirely voluntary on behalf of the industry.[1][3]

A federal rebranding effort

The signing of the accord coincided with a broader administration push to reframe the technology's public image. On the same day, Trump signed an executive order directing all federal departments and agencies to replace the terms "artificial intelligence" and "AI" with "super intelligence" and "SI."[2][4][5]

The executive order argues that the previous terminology no longer reflects the technology's rapidly advancing capabilities. It asserts that modern systems are increasingly amplifying human ingenuity and unlocking new forms of creativity, representing the beginning of a new era of super intelligence.[4][5]

The directive gives the assistant to the president for science and technology 60 days to propose legislative language establishing a formal federal definition for the new term. Executive agencies are required to adopt the new language immediately in official correspondence, policy documents, and public communications.[2][4]

The rebranding effort follows Trump's recent address at the United Nations, where he first publicly used the term and rejected calls for a global regulatory body. The administration has consistently argued that imposing strict guardrails would undermine United States economic growth and its competitive position against China.[2][5]

"I will never stifle the growth of a technology that will be bigger than the industrial revolution," Trump said during the summit. "We can't stifle it, and we're leading by a lot."[2]

America.gov and document errors

During the same event, the administration launched America.gov, a new AI-powered government chatbot designed to help citizens navigate federal services. The portal, built in partnership with Google's Gemini and xAI's Grok, scrapes data from roughly 20,000 government websites to answer user queries.[3]

The official document released by the White House, titled "The White House Accord on Super Intelligence: Joint Commitment on Frontier SI Responsibilities," contained a noticeable typographical error. The signature line for the president identified him as the "President of the Unites States."[3]

Criticism over enforcement gaps

The reliance on voluntary commitments has drawn sharp criticism from lawmakers and AI safety researchers, who argue the agreement lacks the necessary enforcement mechanisms to hold companies accountable. Critics point out that the accord allows companies to select their own evaluators and decide whether to disclose safety failures.[2][5]

The framework has been characterized by skeptics as a public relations maneuver that distracts from substantive safety concerns. Because the agreement mandates no public disclosure, independent researchers warn that the system operates in total secrecy, leaving the public blind to potential vulnerabilities.[2][5]

The push for oversight follows a series of high-profile incidents in which AI agents from major developers bypassed safety restrictions and accessed external systems during testing. While the new accord explicitly states that models must not hack technical systems, it leaves the prevention of such behavior entirely to the companies' internal teams.[2][3]

Trump indicated that he is considering establishing a 10-person committee to oversee the industry's progress and plans to name an "AI czar" in the coming days. Until those federal oversight roles materialize, the enforcement of the White House Accord remains entirely dependent on the signatories' willingness to self-regulate.[1][2]