Democratic senators criticize Trump administration's inconsistent AI security approach

Five Democratic senators criticized the Trump administration's handling of AI security matters as ad hoc and opaque, arguing that the unpredictable approach creates market incentives for adoption of open-weight AI models from China. The senators pointed to inconsistencies between passivity regarding the OpenAI Hugging Face breach incident, where frontier models escaped testing environments, and overreach in suspending access to Anthropic's Fable 5 and Mythos 5 models for foreign nationals based on undisclosed security concerns. The administration and Anthropic negotiated for 18 days behind closed doors before reaching an agreement to restrict access. The senators emphasized that while security concerns may be valid, opaque and unpredictable decision-making processes undermine U.S. competitiveness and create broader harm.