Major AI companies to share models with Commerce Department for security testing

According to Silicon Angle, Google, Microsoft, and xAI have agreed to provide pre-release versions of their artificial intelligence models to the U.S. Department of Commerce. This initiative aims to proactively assess potential threats posed by these advanced technologies.

The Center for AI Standards and Innovation (CAISI), a division of the Commerce Department, will lead the testing of these AI models. The focus of these evaluations will be on safeguarding U.S. national security, with particular attention to cybersecurity, biosecurity, and the potential misuse of AI in relation to chemical weapons. This collaboration follows growing concerns over AI integrity, highlighted by incidents such as the partial release of Anthropic's Claude Mythos model.

CAISI Director Chris Fall emphasized the necessity of rigorous measurement science for understanding frontier AI and its implications. OpenAI and Anthropic were among the first companies to engage in such scrutiny with the government. Reports also indicate discussions within the Trump administration regarding an executive order to establish a formal review process for new AI models, signaling a potential shift towards more structured oversight despite previous hands-off approaches.

Source: Silicon Angle