OpenAI and Anthropic are working together in Washington to form a federal review system for advanced AI models, despite competing for customers, funding and technological leadership.
summary
- OpenAI and Anthropic Support Up to 30 days of federal arrival Before some versions of parametric models.
- The proposed process will be applied across the industry, including competitors e.g Meta and XAI.
- Federal agencies must Develop framework by August 1 By executive order in June.
- Nvidia, Meta, and Microsoft have separately warned about this Comprehensive controls on open-weight AI models.
OpenAI and Anthropic are seeking common review standards
OpenAI and Anthropic are urging the Trump administration to adopt a consistent review process for AI models with advanced cybersecurity or national security capabilities, according to Information.
Their collaboration comes ahead of an August 1 deadline for federal agencies to determine which systems should qualify as “covered border models.” These models can be submitted to the government for evaluation for up to 30 days before being released to other trusted partners.
The two companies reportedly want to apply standards across the AI industry rather than just developers who are already cooperating with Washington. This could put rival companies, including Meta and Elon Musk’s xAI company, under the same review framework if their models exceed the final capability threshold.
The discussions represent a rare policy consensus between OpenAI and Anthropic. Both companies sell general-purpose AI systems and compete for enterprise contracts, researchers, computing power, and investment.
Neither company has publicly released the full terms it wants to include in the final framework.
Trump’s order sets out a 30-day voluntary process
President Donald Trump ordered federal agencies to create a secret measurement process through… Executive Order 14409 On June 2.
The order directs the National Security Agency, the Cybersecurity and Infrastructure Security Agency, the Treasury Department, and other federal agencies to determine when an AI model has capabilities strong enough to warrant additional review.
Developers participating in the framework will be able to ask the government whether the model meets the threshold for a covered boundary model. They can then provide federal residents with access for up to 30 days before making the system available to other authorized partners.
Intellectual property, confidential information, cybersecurity and insider threat safeguards are required during the assessment.
However, it explicitly states that the process cannot create a mandatory federal license, prior authorization, or rules permitting the development or release of AI models. Therefore, companies will participate voluntarily unless separate laws or government powers apply.
US AI developers seek clarity ahead of launch
The common framework could give US developers clearer criteria for determining when to expect government testing. Questions remain about what technical standards will trigger the review, which agencies will test each system, and how officials will approve early access partners.
These uncertainties have already affected recent launches. OpenAI and Anthropic have provided government officials with early access to advanced models before wider distribution, but outside researchers have had limited visibility into the evaluations.
A July report It found that the agencies, testing and approval requirements involved in those reviews remained unclear. The lack of a published process can make release planning more difficult for developers, and leave small AI companies unsure if the same standards apply to them.
Applying a single capabilities-based threshold to OpenAI, Anthropic, Meta, xAI, and other developers could reduce some of this uncertainty. It could also reduce concerns that Washington is giving select companies different treatment.
The open debate about artificial intelligence is dividing the industry
OpenAI and Anthropic’s call for a joint review process comes as other tech companies warn Washington against imposing broad controls on open-weight AI models.
Such as crypto.news I mentioned Last week, Nvidia, Meta and Microsoft joined 22 other organizations in an open letter arguing that sweeping restrictions could weaken US leadership as competition with China intensifies. IBM, Palantir, Mistral, Hugging Face, Mozilla, Andreessen Horowitz, and the Linux Foundation also signed the letter.
The group called for targeted legal and commercial action against misuse rather than restrictions covering technologies with legitimate research and business applications.
Open-weight models allow companies, researchers, and governments to download, modify, and run software on their own infrastructure. Proponents say this gives users greater control over their data, security, and computing systems while lowering deployment costs.
The signatories did not position open and closed models as mutually exclusive. Instead, they argued that both approaches are necessary to support competition and give developers different ways to build and deploy AI systems.
Nvidia CEO Jensen Huang shared the message in his first post on X and defended the role of both development models.
“The world needs closed border models and open border models.”
Elon Musk also endorsed the message in his response to Huang, although his artificial intelligence company, xAI, was not listed among the 25 sites reported.
These two efforts address different parts of the political debate. OpenAI and Anthropic seek consistent federal evaluations of models that exceed the national security capability threshold, while the open letter argues against restrictions that rely primarily on whether a model makes its weights available.
Federal agencies now have until August 1 to set the threshold for the border model and determine a voluntary review process. The final framework will clarify whether Washington focuses on measurable capabilities or applies broader conditions based on how AI models are distributed.
