The Trump administration has finalised a framework this week for testing new artificial intelligence models for safety and cybersecurity risks. However, the White House is keeping the details of this framework private.
On Tuesday, representatives from OpenAI, Anthropic, Meta, Google, Nvidia, and Microsoft attended a private meeting with White House officials to review the AI framework. Multiple outlets have since reported that while the volunteer vetting process for new AI models has been settled, the White House does not plan to release its policy publicly. Instead, it will only share testing criteria with a select few tech companies.
This opaque process has raised questions about how the framework will operate, including the level of scrutiny models will face and the safety benchmarks they must meet. Businesses, foreign governments, and the public currently lack clarity on these details.
Discussions for an AI cybersecurity framework began earlier this year following concerns over Anthropic's Mythos model, which was withheld from public release due to potential misuse for hacking IT and financial systems. In June, the White House issued an executive order calling for AI companies to voluntarily submit new models for government review up to 30 days before release.
The executive order does not define what qualifies as the advanced AI it targets, leaving it unclear which companies' models will be subject to government review. Open-source models, for example, will be excluded from the framework, according to Axios.