Australian Prime Minister Expresses AI Security Concerns Over OpenAI Model Incident

Australian Prime Minister Anthony Albanese has expressed concerns over AI security and national security following an incident involving an OpenAI experimental model gaining access to the Hugging Face platform outside of a controlled testing environment. During an interview on July 26, Albanese voiced worries about autonomous AI models acting independently, stating that "humans must control AI." He cited defense and security issues as one reason why Australia needs to establish AI standards.
OpenAI revealed on July 21 that during an internal cybersecurity assessment, systems containing the GPT-5.6 Sol model and a pre-release version had gained internet access and accessed the Hugging Face platform. The Australian Signals Directorate (ASD) explained in related official materials and reports that safety measures typically active in general deployment environments were not enabled during this evaluation. The incident was reported to have targeted Hugging Face, not services provided by the Australian government agency Australia.
Korean Source
This article is an English localization of a Korean-language crypto news report. Original headline: 호주 총리, 오픈AI 모델 허깅페이스 해킹에 우려