Alphabet’s Google, Microsoft, and xAI have entered an agreement with the Trump administration to share early versions of their artificial intelligence models with federal officials for security assessments prior to public release. The arrangement is intended to enable a thorough evaluation of the AI systems' capabilities and potential risks.

The U.S. Commerce Department’s Center for AI Standards and Innovation (CAISI) will oversee these evaluations. CAISI has previously conducted more than 40 assessments, including of AI models that have not yet been made public. This follows a similar arrangement established in 2024 with ChatGPT developer OpenAI and AI company Anthropic.

“Independent, rigorous measurement science is essential to understanding frontier AI and its national security implications,” CAISI Director Chris Fall said on Tuesday. He added that the expanded collaborations with industry are designed to enhance the government’s capacity to analyze emerging AI technologies in the public interest.

Under the terms of the agreement, AI developers will provide models to CAISI with reduced or removed safety safeguards so that the department can evaluate possible national security risks and capabilities without limitations imposed by built-in protections.

This cooperative framework is part of broader U.S. government considerations regarding the implementation of a review process for artificial intelligence tools that may present cybersecurity threats. The administration is exploring mechanisms to identify and mitigate risks posed by advanced AI before those technologies reach commercial markets.

Notably, News Corp—the owner of the reporting outlet—maintains commercial and content-licensing partnerships with Google and OpenAI, respectively. However, the current agreements reflect ongoing efforts to balance AI innovation with national security vigilance.