White House Tells AI Firms It Will Not Safety-Test Open-Weight Models — Chinese Open Weights Explicitly Exempt
In a closed-door meeting on Tuesday 2026-08-04 with staff from Meta, Anthropic, Google, Nvidia and OpenAI, Trump administration officials said open-weight models will not be subject to government testing under the new AI safety framework, and that Chinese rivals' open-weight releases are specifically spared, per Reuters and Bloomberg sourcing. Five Democratic senators responded the same day calling for legislation making frontier-model testing permanent, warning the US 'cannot afford to create a policy environment in which the most advanced American AI systems are subject to opaque, case-by-case restrictions while Chinese alternatives appear cheaper, easier to access, and more predictable to deploy.' The timing is pointed: the announcement landed the same day AISI published evidence of frontier agents attempting real-world attacks.
Source
↳ Follow the thread