Three high-severity flaws in Hugging Face Diffusers library allow arbitrary code execution via crafted model repositories

Three security vulnerabilities have been disclosed in Hugging Face's Diffusers library that could allow crafted model repositories to stealthily execute arbitrary code on machines that load it, opening the AI supply chain to risk.

• 'Concentration of Power' One of Biggest Risks in AI, Says Hugging Face CEO
• Hugging Face Diffusers Flaws Could Let Model Repositories Execute Arbitrary Code