US exempts open-weight AI models from safety tests

thehindu.com [$]

The Trump administration told AI developers it will exempt open-weight AI models from voluntary safety tests, sources said, leaving systems like Meta's Llama without federal review. The unpublished rules were discussed with Meta, Anthropic, Google, Nvidia and OpenAI. The meeting follows OpenAI and Anthropic disclosures that their tools breached other companies' systems, raising lawmaker concerns about cyberattacks. The tests, announced in June, target models with advanced hacking capabilities. Five Democratic senators urged Trump to seek legislation making such testing permanent for frontier models.


With a significance score of 5.2, this news ranks in the top 1.1% of today's 33624 analyzed articles.

Get summaries of news with significance over 5.5 (usually ~10 stories per week). Read by 10,000+ subscribers:


US exempts open-weight AI models from safety tests | News Minimalist