New executive order asks OpenAI, Anthropic, Google and other AI developers to voluntarily give the federal government a 30-day advance look at powerful new models before public release.

By JBizNews Desk

June 3, 2026

President Donald Trump on Tuesday signed an executive order titled “Promoting Advanced Artificial Intelligence Innovation and Security,” asking the country’s leading artificial-intelligence companies to provide the federal government with early access to their most advanced AI systems before they are released more broadly.

The order, published by the White House, establishes a voluntary framework under which AI developers would share powerful new models with government reviewers up to 30 days before providing them to outside partners or wider markets.

Importantly, the administration emphasized that the program is voluntary.

The order specifically states that companies will not be required to obtain government approval, licenses, or pre-clearance before launching new AI products.

A Voluntary Approach, Not a Licensing System

That distinction appears designed to reassure the AI industry.

Rather than imposing a new regulatory gatekeeper, the administration is asking companies such as OpenAI, Anthropic, and Google to cooperate voluntarily with government reviewers who would evaluate security and national-security implications before broader deployment.

The White House framed the initiative as a way to balance innovation with security.

“Advanced AI capabilities make our Nation stronger, but also introduce new national security considerations,” the executive order states.

Federal agencies participating in the effort include the Department of Defense, the Treasury Department, and the Cybersecurity and Infrastructure Security Agency (CISA).

The goal is to identify vulnerabilities that could affect critical infrastructure, including electric grids, financial systems, communications networks, and water utilities.

The Order Nearly Didn’t Happen

Tuesday’s signing followed weeks of internal debate.

The administration originally planned to unveil the order in late May during a White House event attended by major technology executives. That rollout was abruptly canceled hours before the scheduled announcement.

At the time, President Trump expressed concerns that excessive oversight could weaken America’s competitive advantage.

“We’re leading China, we’re leading everybody,” Trump told reporters.

Industry leaders had also objected to an earlier draft that reportedly included a 90-day review period. The final version cuts that timeline to 30 days, reflecting concerns that AI technology now evolves too rapidly for longer review windows.

Reports indicated that former White House AI adviser David Sacks advocated scrapping the original proposal and replacing it with a lighter-touch framework.

What Triggered the Push

The administration’s concern accelerated after the release of Anthropic’s Mythos Preview model in April.

According to officials familiar with the matter, the model demonstrated an ability to identify serious vulnerabilities in widely used software systems at a speed that alarmed cybersecurity officials.

The same capability could help defend critical systems—or potentially help attackers exploit them.

That development reportedly triggered months of discussion inside Washington about how government agencies could gain visibility into cutting-edge AI systems without imposing regulations that might slow innovation.

Senior Officials Take a Larger Role

The executive order reflects growing involvement from senior administration officials.

Treasury Secretary Scott Bessent and Vice President JD Vance have both become increasingly active in discussions surrounding AI’s economic and national-security implications.

The effort also follows earlier signals from the National Institute of Standards and Technology (NIST) that major technology companies would voluntarily share advanced models with government reviewers for safety testing.

A Softer Outcome Than Many Expected

For AI companies, the final order is significantly less restrictive than many had feared.

There is:

• No licensing requirement
• No mandatory approval process
• No government authorization needed before release
• No formal pre-launch certification system

Those are precisely the types of regulations that AI developers have argued could hinder America’s ability to compete globally.

Some elements of the order simply formalize practices already emerging within the industry. OpenAI, for example, launched its own voluntary early-access testing program earlier this year.

Many technology executives view cooperation as a way to avoid more aggressive regulation later.

A Complicated Relationship With Industry

The order arrives amid continuing tensions between Washington and some AI developers.

The administration remains involved in legal disputes with Anthropic stemming from disagreements over military applications of AI technology.

Anthropic has maintained restrictions preventing its systems from being used for lethal autonomous weapons or large-scale domestic surveillance operations.

That position led to friction with portions of the defense establishment and remains the subject of ongoing litigation.

Why Businesses Should Care

The implications extend far beyond Silicon Valley.

The same advanced AI models capable of identifying vulnerabilities in software can also strengthen cybersecurity defenses.

Banks, utilities, hospitals, telecommunications providers, and infrastructure operators all have a direct stake in how these systems are evaluated before deployment.

A flaw discovered before release could prevent widespread damage.

A flaw missed until after release could create significant risks.

The Bottom Line

President Trump’s new executive order attempts to strike a middle ground between oversight and innovation.

Rather than requiring government approval, it asks AI companies to voluntarily provide an early look at their most powerful systems before public release.

For now, Washington is relying on cooperation rather than regulation.

Whether the largest AI labs continue participating voluntarily as the technology grows more powerful may determine whether future administrations pursue a much stricter approach.

Washington — JBizNews Desk

© JBizNews.com All Rights Reserved. Reproduction or distribution without written permission is prohibited.