Copyright Violations and Large Language Models
This addresses the issue of copyright compliance for developers and users of natural language processing models, but it is incremental as it builds on existing concerns about memorization.
This work investigates the problem of copyright violations in large language models due to verbatim memorization of copyrighted texts, such as books and coding problems, and provides a conservative characterization of the extent of redistribution, though no concrete numbers are reported.
Language models may memorize more than just facts, including entire chunks of texts seen during training. Fair use exemptions to copyright laws typically allow for limited use of copyrighted material without permission from the copyright holder, but typically for extraction of information from copyrighted materials, rather than {\em verbatim} reproduction. This work explores the issue of copyright violations and large language models through the lens of verbatim memorization, focusing on possible redistribution of copyrighted text. We present experiments with a range of language models over a collection of popular books and coding problems, providing a conservative characterization of the extent to which language models can redistribute these materials. Overall, this research highlights the need for further examination and the potential impact on future developments in natural language processing to ensure adherence to copyright regulations. Code is at \url{https://github.com/coastalcph/CopyrightLLMs}.