Hugging Face Used Chinese AI Model to Probe Autonomous Cyberattack

Hugging Face said it used a Chinese open-source AI model, GLM 5.2 from Z.ai, to investigate and contain a fully autonomous cyberattack after safety filters in U.S. proprietary models blocked its security team from analyzing the attack data. The company said those guardrails could not tell an incident responder from an attacker, forcing it to keep the analysis on its own infrastructure and examine more than 17,000 logs and traces left by the intrusion. The episode has sharpened the debate over whether strict AI safety controls can hinder legitimate cybersecurity work, especially when defenders need fast access to raw attack payloads and forensic details.