Summary
A recent Hugging Face cybersecurity incident revealed that a Chinese open-source AI model, GLM-5.2, successfully aided in investigation where American models failed due to safety guardrails. This event has intensified the US policy debate over open-weight AI, with Washington considering restrictions while Silicon Valley companies warn against them.