US government to test AI models before they are released

On 2026-05-07, the U.S. Commerce Department’s National Institute of Standards and Technology said its Centre for AI Standards and Innovation would conduct pre-deployment evaluations of new AI models from Google, Microsoft, and xAI before release. The arrangement expands government oversight of frontier AI systems and builds on a prior prerelease testing deal with OpenAI and Anthropic. It is an official same-day policy action affecting model deployment oversight.