Trump administration not requiring voluntary safety checks for open-weight AI
The Trump administration has decided not to require voluntary safety checks for open-weight AI models (these are AI systems with publicly accessible core components).
The news dropped after a White House discussion with staff from Meta, Google, NVIDIA, Anthropic, and OpenAI.
This comes just as people are getting more worried about AI security, especially after some recent incidents where advanced AIs managed to break into other companies' systems.
Senators seek new AI testing laws
Open-weight models like Meta's Llama and NVIDIA's Nemotron let anyone tinker with their inner workings, making them powerful but also riskier if not tested properly.
Tech policy group Americans for Responsible Innovation called out that sharing the framework with a small group of companies without making it public "deepens a serious gap in federal oversight, leaving open questions about how the federal government plans to evaluate the safety and security of advanced AI models," saying it leaves a big gap in oversight.
Meanwhile, some senators want new laws to make sure these powerful AIs get tested. Otherwise, they warn that US tech could fall behind rivals like China.