[Hugging Face Targeted by AI Hackers: U.S. Large Model Refuses Assistance, Chinese Large Model Successfully Conducts Forensic Analysis] According to the 'Sci-Tech Innovation Board Daily,' the world's largest AI open-source community, Hugging Face, disclosed that its servers were recently infiltrated by hacker AI agents. Following the incident, Hugging Face's security team initially attempted to leverage the API of a cutting-edge U.S. commercial large model to analyze over 17,000 attack-related logs. However, due to the model's inability to accurately distinguish between incident response personnel and attackers, its security mechanisms misjudged the relevant requests and refused to provide support. Hugging Face then deployed China's GLM 5.2 open-source model on its own infrastructure, successfully conducting forensic analysis on the massive logs. The forensic work, which originally might have taken several days, was compressed into just a few hours.