Search papers, labs, and topics across Lattice.
This study investigates the propagation of licensing obligations in AI supply chains by analyzing 232,270 dataset-to-model-to-application chains. The findings reveal that a staggering 62.3% of these chains include at least one artifact lacking a declared license, while the survival rate of obligation-bearing licenses is alarmingly low, with none exceeding 7% compared to a 95.1% survival rate for permissive licenses. These results highlight significant vulnerabilities in the licensing framework of AI artifacts, prompting actionable recommendations for stakeholders in the ecosystem.
Over 62% of AI supply chains include artifacts with no declared license, raising serious concerns about compliance and accountability in AI development.
AI artifacts move through a multi-platform supply chain, spanning datasets and models on Hugging Face and applications on GitHub. While each artifact carries a license whose obligations should propagate through redistribution, no study has yet measured whether those obligations survive the chain or are stripped and replaced as artifacts move downstream. We trace 232,270 dataset$\rightarrow$model$\rightarrow$application chains and quantify two forms of license laundering: when artifacts with no declared license acquire definitive labels downstream, and when one declared license category replaces another during redistribution. We find that 62.3% of chains pass through at least one artifact with no declared license (concentrated in a small set of foundational datasets), and that every obligation-bearing license category falls below 7% end-to-end survival while the Permissive category reaches 95.1%. Based on these findings, we provide actionable recommendations for practitioners, model publishers, rights holders, and platform owners.
