Hugging Face Forced to Pivot to Open-Weights Model During AI Attack
Hugging Face had to swap to Z.ai's open-weights GLM 5.2 after commercial guardrails refused to engage an attacker — exposing a real agentic AI security gap.
Hugging Face had to swap to Z.ai's open-weights GLM 5.2 after commercial guardrails refused to engage an attacker — exposing a real agentic AI security gap.