Microsoft, Google, xAI to Open Next-Gen AI Models to U.S. Government for Security Review Before Launch

As cutting-edge AI technology develops rapidly, national security risks have attracted attention. Recently, Microsoft, Google, and xAI founded by Elon Musk have reached a new agreement with the U.S. government to open access to their next-generation AI models before official launch.

The U.S. Department of Commerce's AI Standards and Innovation Center stated that the agreement allows pre-deployment assessment of model capabilities and potential risks such as cyberattacks and military misuse. This expands the 2024 Biden administration agreement, and the center has completed over 40 evaluations of advanced models so far.

In addition, the U.S. Department of Defense recently partnered with seven AI companies to deploy technology to classified networks, but Anthropic was not included due to permission issues.