“Independent, rigorous measurement science is essential to understanding frontier AI and its national security implications,” CAISI Director Chris Fall said in a statement.Microsoft, Alphabet-owned Google and Elon Musk’s xAI have agreed to provide the US government with early access to new artificial intelligence (AI) models before their public release, allowing officials to assess potential national security risks, according to Reuters.
The Center for AI Standards and Innovation (CAISI) at the US Department of Commerce said on Tuesday the arrangement would allow it to evaluate models before deployment and conduct research into their capabilities and security risks.
The agreement expands the government’s oversight of advanced AI systems amid growing concerns in Washington about the risks posed by increasingly powerful models, including their potential use in cyberattacks and military applications.
“Independent, rigorous measurement science is essential to understanding frontier AI and its national security implications,” CAISI Director Chris Fall said in a statement.
AI scrutiny expandsThe latest arrangement builds on similar agreements signed with OpenAI and Anthropic in 2024 under the Biden administration, when CAISI operated as the US Artificial Intelligence Safety Institute.
CAISI said it has already completed more than 40 evaluations, including assessments of advanced AI models that had not yet been released publicly.
According to the agency, developers often provide versions of their models with safety guardrails reduced or removed to allow the centre to test for national security vulnerabilities.
The development of advanced AI systems, including Anthropic’s Mythos, has recently drawn attention from US officials and corporate America over concerns that such tools could strengthen hacking capabilities.
The announcement follows a Pentagon decision last week to reach agreements with seven AI companies to deploy advanced AI capabilities across the Defense Department’s classified networks.
Anthropic was not included in those agreements and has been involved in a dispute with the Pentagon over safeguards governing the military use of its AI systems.
Published On May 6, 2026 at 12:19 PM IST
Join the community of 2M+ industry professionals.
Subscribe to Newsletter to get latest insights & analysis in your inbox.
Get updates on your preferred social platform
Follow us for the latest news, insider access to events and more.