With these agreements, the AI developers align themselves with OpenAI and Anthropic PBC in permitting pre-release model reviews by the US Commerce Department’s Centre for AI Standards and Innovation, as announced by the agency on Tuesday. OpenAI and Anthropic have updated their existing partnerships with the centre to ensure better alignment with priorities outlined in President Donald Trump’s AI Action Plan, according to the agency.
Since 2024, the office has been evaluating models from OpenAI and Anthropic before they are released to the public. The agency has conducted over 40 evaluations of AI models, including cutting-edge models that are yet to be made public, as stated.
These agreements are being revealed as Anthropic’s Mythos system raises concerns among US officials, indicating a broader mandate for the relatively new centre, which was established under President Joe Biden as the AI Safety Institute in 2023 and rebranded by the Trump administration last year.
The centre positions itself as the “industry’s primary point of contact within the US government” for testing, collaborative research, and the development of best practices. Although its existence has not been codified into law, some lawmakers have introduced draft legislation to provide the centre with a more permanent framework.
“These expanded industry collaborations assist us in scaling our public interest work at a critical time,” said Chris Fall, the centre’s director. Fall took over following the unexpected exit of Collin Burns, a former AI researcher at Anthropic, who was selected for the role but left just days after starting, according to reports.
Also Read: Larsen and Toubro wins ‘mega’ order from JSW Steel, its largest ever in domestic metals sector
The new evaluation agreements come after reports from the New York Times and the Wall Street Journal indicating that the Trump administration is contemplating an executive order to establish a government review process for AI tools. This review process would serve as a form of oversight, the reports indicate. A White House official stated that any announcement would originate from Trump and characterized discussions around possible executive orders as speculation.
Trump’s AI Action Plan, released in July, mandates that the Centre for AI Standards and Innovation be involved in a so-called AI evaluations ecosystem and lead assessments related to national security AI models.
The plan further suggests that regulators should “explore the use of evaluations in their applications of existing law to AI systems.” The centre’s broader evaluation of models might facilitate new enforcement of currently existing laws.
The administration’s efforts to influence AI policy have intensified since Anthropic disclosed last month that its groundbreaking Mythos model was skilled at identifying vulnerabilities in cybersecurity defenses.
White House Chief of Staff Susie Wiles, Treasury Secretary Scott Bessent, and National Cyber Director Sean Cairncross have increasingly engaged in matters related to Mythos, while the White House has opposed Anthropic’s proposal to expand access to its Mythos model.
Any Commerce Department evaluation agreement with Anthropic could be complicated by the Pentagon’s ongoing legal disputes with the company, involving two lawsuits about whether the Defence Department can officially classify the AI firm as a supply-chain risk.
Both Defence Secretary Pete Hegseth and Trump have called for a six-month phase-out for government reliance on Anthropic’s tools, while an upcoming White House memo regarding agencies’ AI usage will address tense points in contract negotiations from both perspectives.
