Hugging Face says it resorted to a Chinese AI model to battle a fully autonomous cyberattack because U.S. model guardrails stymied its defense

Hugging Face, a platform for hosting and collaborating on machine learning models, has disclosed that it turned to a Chinese AI model to defend against a fully autonomous cyberattack. The attack was described as fully autonomous, meaning it operated without human intervention.

The decision was made after the US model's safety features hindered its ability to assist in the defensive operation. The incident highlights the complexities of using AI systems with built-in restrictions during security incidents. Hugging Face did not name the US model that was unable to help.

The event underscores potential limitations of AI guardrails—mechanisms designed to prevent misuse—when applied to defensive cybersecurity scenarios. It also demonstrates the global availability of AI models, with organizations potentially turning to less restricted alternatives from other countries. The use of a Chinese model for defense against an autonomous attack raises considerations about supply chain dependencies in AI security.