LicenseGPT: A Fine-tuned Foundation Model for Publicly Available Dataset License Compliance
Jingwen Tan, Gopi Krishnan Rajbahadur, Zi Li, Xiangfu Song, Jianshan Lin, Dan Li, Zibin Zheng, Ahmed E. Hassan
TL;DR
This work introduces LicenseGPT, a fine-tuned foundation model tailored to dataset license compliance analysis. By building and annotating a Dataset Licenses (DL) corpus of 500 licenses and using a structured prompt design, the authors show LicenseGPT achieves $PA = 64.30\%$ and $SS = 85.80\%$, outperforming existing legal FMs and general-purpose models in accuracy while maintaining practical response times. Through a zero-shot, 10-fold cross-validation study and practitioner-focused A/B tests with software IP lawyers, the paper demonstrates substantial efficiency gains (up to $94.44\%$ time reduction) and positive reception, albeit with acknowledged need for human oversight in nuanced cases. The work also provides open-source access to LicenseGPT and offers recommendations to integrate automated license analysis into AI development lifecycles, along with future research directions around data-size effects and standardization of license metadata.
Abstract
Dataset license compliance is a critical yet complex aspect of developing commercial AI products, particularly with the increasing use of publicly available datasets. Ambiguities in dataset licenses pose significant legal risks, making it challenging even for software IP lawyers to accurately interpret rights and obligations. In this paper, we introduce LicenseGPT, a fine-tuned foundation model (FM) specifically designed for dataset license compliance analysis. We first evaluate existing legal FMs (i.e., FMs specialized in understanding and processing legal texts) and find that the best-performing model achieves a Prediction Agreement (PA) of only 43.75%. LicenseGPT, fine-tuned on a curated dataset of 500 licenses annotated by legal experts, significantly improves PA to 64.30%, outperforming both legal and general-purpose FMs. Through an A/B test and user study with software IP lawyers, we demonstrate that LicenseGPT reduces analysis time by 94.44%, from 108 seconds to 6 seconds per license, without compromising accuracy. Software IP lawyers perceive LicenseGPT as a valuable supplementary tool that enhances efficiency while acknowledging the need for human oversight in complex cases. Our work underscores the potential of specialized AI tools in legal practice and offers a publicly available resource for practitioners and researchers.