OpenAI to Give Government Early Access to New AI Models Under Trump Order

Benzinga reported Friday that OpenAI intends to join a voluntary White House program giving federal agencies government early access to AI models ahead of their public launch.

OpenAI Commits to the Trump Pre-Release Framework

The program stems from an executive order signed by President Donald Trump. It asks participating companies to hand over access to new AI systems 30 days before release. During that window, government evaluators run standardized benchmarking tests. Those tests are designed to measure advanced capabilities, including potential cyber-related risks embedded in the models.

OpenAI executive George Osborne told CNBC the company plans to enroll in the scheme. He framed the move as part of a broader philosophy around democratic oversight of transformative technology. Osborne argued that elected governments have a legitimate and necessary role in shaping how AI is deployed at scale. He also indicated OpenAI prefers to bring ideas to regulators proactively rather than simply waiting to receive mandates from above.

A Push for Adaptive Regulatory Bodies

Osborne went further, calling on governments worldwide to establish well-resourced regulatory institutions capable of keeping pace with rapid AI development. The argument reflects a growing consensus among major AI developers that voluntary cooperation now may forestall heavier-handed legislation later.

Separately, OpenAI published a policy blog post clarifying its stance on political advocacy. The company stated plainly that no external political organization speaks on its behalf. It also said AI policy is too consequential to become a partisan battleground, urging that the technology’s governance remain above standard political division.

Background: OpenAI and Washington’s Evolving Relationship

OpenAI has maintained a consistent presence in Washington since the late-2022 launch of ChatGPT turbocharged public and legislative interest in generative AI. Chief Executive Sam Altman has testified before Congress on multiple occasions. Those hearings helped put AI governance firmly on the federal agenda. The current pre-release access program represents a more structured, institutionalized version of the informal dialogue that has been building for several years.

The framework is voluntary for now. That detail matters. Critics argue voluntary programs lack enforcement teeth. Supporters counter that early cooperation builds the regulatory muscle memory needed before mandatory rules arrive. OpenAI’s participation lends the initiative significant credibility given the company’s market position and public profile.

Whether other frontier AI developers follow OpenAI’s lead will determine how meaningful the program becomes in practice.

Read Next: AI Regulation Heats Up as Governments Race to Keep Pace With Technology