Hugging Face Turns to Chinese AI Model After Cyberattack Highlights U.S. Safety Limits

Hugging Face turned to Chinese AI startup Zhipu AI's open-source GLM-5.2 model to analyze an intrusion last week after major U.S. AI models refused to help due to built-in safety restrictions. The New York startup had requested assistance from leading American models including Anthropic's Claude and OpenAI's GPT series, but these systems blocked cybersecurity-related queries to prevent misuse, making them unable to distinguish defenders from attackers.

The incident highlights a growing tension between safety guardrails and practical capability. Hugging Face co-founder Clement Delangue stated that enterprises are learning that secrecy is not the solution; effective defense requires more powerful, less-restricted models. Zhipu's GLM-5.2, which achieved parity with OpenAI and Anthropic in coding and agent capabilities while offering lower costs, has rapidly gained attention among Silicon Valley developers since its launch last month.