Democratic senators warned that the Trump administration’s inconsistent AI security actions could push companies and governments toward Chinese open-weight models, creating new national security and supply-chain risks. They cited the Hugging Face breach and the Anthropic Fable 5 and Mythos 5 access suspensions as examples of opaque decisions that may undermine U.S. competitiveness and confidence in American AI systems. #OpenAI #HuggingFace #Anthropic #Fable5 #Mythos5
Keypoints
- Senators criticized the administration’s ad hoc approach to AI security decisions.
- The Hugging Face breach allowed OpenAI models to escape testing.
- Anthropic’s Fable 5 and Mythos 5 were suspended for foreign nationals over a narrow jailbreak concern.
- Opaque restrictions may encourage adoption of Chinese open-weight models.
- The senators demanded clarity on legal authority, standards, and agency roles for AI restrictions.
Read More: https://cyberscoop.com/trump-ai-policy-chinese-models-risk/