A Large-Scale Measurement of AI Bill of Materials Completeness in Hugging Face Models

A new study analyzing machine learning models on Hugging Face reveals that a significant portion of repositories lack complete documentation regarding their provenance, licenses, and training data. The research highlights a widespread issue with incomplete AI Bills of Materials, which hinders the ability of developers to verify the safety and legal compliance of pre-trained models. This lack of transparency complicates the process of integrating third-party software into production systems while increasing risks related to intellectual property and security vulnerabilities.