White House officials are set to expand their new artificial intelligence guidelines, preparing to subject even more powerful models to federal safety tests before release.

The move reflects growing concerns over potential security risks and the need for real-time adaptation as AI technologies rapidly evolve. The framework, which currently covers only closed models like Anthropic's Mythos-class or OpenAI’s GPT-5.6, is expected to include open models once they match similar capabilities.

While this approach aims to ensure public safety, it could also create a two-tier system where enterprises might hesitate to use non-approved open models despite their cost benefits. The concern lies in the potential for these models to autonomously hack critical systems like the Pentagon or financial markets.

In a twist of irony, while President Trump initially resisted formal regulation fearing China's advancement, pressure from other parts of his administration is pushing for more robust guidelines with leading AI labs.