US to Review AI Models from Microsoft, Google, xAI
The National Institute of Standards and Technology announced on Tuesday that Google, Microsoft, and xAI will provide the government with unreleased versions of their AI models to address cybersecurity threats.
The collaboration follows Anthropic's robust new Mythos AI model that heightened worries regarding AI's influence on cybersecurity last month, leading the White House to consider an official evaluation process for AI.
The updated agreements empower the Center for AI Standards and Innovation, part of the US Department of Commerce, to assess emerging AI models and their possible effects on national security and public safety before they are released. The center will additionally perform research and testing following the deployment of AI models and has already finished over 40 evaluations of AI models.
“Independent, thorough measurement science is crucial for grasping frontier AI and its implications for national security,” stated CAISI Director Chris Fall. “Such broadened industry partnerships enable us to amplify our efforts for the public good during a crucial time.”
Mythos, described by Anthropic as being "well in advance" of other models regarding cybersecurity, raised significant concerns among governments, banks, and utility companies in the last month. The firm stated it is not ready to share the model with the public and is limiting access to a specific set of authorized organizations. It has also updated high-ranking US government officials on its abilities.
Also Read: How This Techie Turned Visa Struggles into Startup Success
Last week, OpenAI announced that it is providing its most advanced AI models to all verified government levels to anticipate AI-related threats.
Jessica Ji, a senior research analyst at Georgetown’s Center for Security and Emerging Technology, stated that the collaborations could facilitate CAISI's AI testing by offering additional resources.
Also Read: 5 Key Leadership Appointments across Global Firms in March 2026
“They just lack the same resources (as major tech firms), including workforce, technical personnel, and access to computing power, to refine these models and conduct thorough testing,” she stated.
Also Read: How US–China Tariffs Are Disrupting Textile Trade
The White House is seeking to consult with a panel of specialists to provide guidance on a potential governmental review procedure for emerging AI models, as confirmed by CNN. This would mark a shift from the Trump administration's lenient stance on AI regulation up to this point.