Whose Text Is It Anyway? Exploring BigCode, Intellectual Property, and Ethics

Abstract: Intelligent or generative writing tools rely on LLMs that recognize, summarize, translate, and predict content. This position paper probes the copyright interests of open data sets used to train LLMs. Our paper asks, how do LLMs trained on open data sets circumvent the copyright interests of the used data? We start by defining software copyright and tracing its history. We rely on GitHub Copilot as a modern case study challenging software copyright. Our conclusion outlines obstacles that generative writing assistants create for copyright, and offers a practical road map for copyright analysis for developers, software law experts, and general users to consider in the context of intelligent LLM-powered writing tools.