Generative AI and Fair Use: A ChatGPT Case Study

This article examines whether generative AI models, particularly ChatGPT, qualify for fair use of copyrighted material. The author analyzes the four factors outlined in Section 107 of the US Copyright Act: purpose and character of use, nature of the copyrighted work, amount and substantiality of the portion used, and effect on the market. Through a case study of ChatGPT, the author argues that ChatGPT's use of its training data likely constitutes copyright infringement. ChatGPT's commercial nature and its failure to transform the training data, coupled with market harm to original works, contradict the principles of fair use.