The Trump administration has made agreements with Google DeepMind, Microsoft, and xAI to evaluate artificial intelligence models before their release.

The U.S. administration announced on Tuesday that it has reached agreements with Google DeepMind, Microsoft, and xAI to gain access to the most advanced artificial intelligence models, which will be evaluated before their release. This represents a significant shift in the AI regulation approach by the Trump administration, which had previously been reluctant to address this issue. The recent partnerships are inspired by similar agreements from the tenure of former President Joe Biden, although Trump claims that they have been renegotiated.

The White House is working on an executive order to create a working group made up of government officials and leaders from the tech sector, which will analyze new AI models. The Center for AI Standards and Innovation (CAISI) will conduct assessments to better understand the advanced capabilities of AI and to promote its security. CAISI Director Chris Fall emphasized the importance of rigorous scientific measures for national security. Additionally, agencies like the NSA are already testing new AI models, such as Mythos, to identify security vulnerabilities.