OpenAI Complies With Trump AI Order

OpenAI said in an on-air interview on June 5, 2026, that it will comply with President Trump’s AI executive order and join the voluntary pre-release review, allowing U.S. agencies to assess certain models up to 30 days before public release.

OpenAI Commits to Voluntary Pre-Release Review

George Osborne, OpenAI’s head for countries, said the company will “sign up to the voluntary framework established by the order.” This aligns OpenAI with the administration’s national-security approach to high-capability AI systems. By subjecting model releases to government assessment, OpenAI may adjust the timing and operational rollout of those systems.

Trump Order Establishes 30-Day Review and Agency Roles

President Trump signed the executive order titled “Promoting Advanced Artificial Intelligence Innovation and Security” on June 2, 2026. The order creates a voluntary federal framework to address cybersecurity risks from advanced AI systems described as frontier models. It sets a 30-day pre-deployment review window, shorter than an earlier draft that proposed a longer period.

The order authorizes the National Security Agency director to develop a classified benchmarking process to identify systems that exceed capability thresholds and qualify as “covered frontier models.” It assigns Treasury, Defense (through the NSA), Homeland Security (through the Cybersecurity and Infrastructure Security Agency, CISA), the White House science office, and Commerce (through the National Institute of Standards and Technology, NIST) to build the operational framework. CISA is instructed to issue binding operational directives under the Federal Information Security Modernization Act to strengthen federal civilian cybersecurity defenses.

The voluntary assessments focus on advanced cyber capabilities, including identifying and exploiting software vulnerabilities—a risk flagged in recent evaluations of Anthropic’s Mythos and OpenAI’s GPT 5.5. The pre-deployment testing aims to detect novel behaviors and determine whether a model should be designated a “covered frontier model,” a status that could restrict its distribution and sale.

The order also directs Treasury to establish an AI cybersecurity clearinghouse within 30 days to coordinate vulnerability scanning and responses among AI labs and critical infrastructure operators. The Office of Management and Budget has 60 days to evaluate available funding for this effort.

Officials Discuss Financial Stakes in AI Firms

U.S. officials have discussed taking financial stakes or structured exposure in AI companies as part of a broader national AI strategy. These talks have involved leading developers, including OpenAI CEO Sam Altman, who proposed the idea. The discussions considered instruments such as equity or structured stakes that could give the government financial exposure or influence in key AI firms.

Together, OpenAI’s public commitment to the voluntary review and the administration’s framework—including classified benchmarks, binding directives, and potential financial levers—signal closer alignment between frontier AI labs and national-security agencies. This may prompt firms to adjust release practices, governance, and testing locations as high-capability models emerge.