Users’ perceptions of generative artificial intelligence and copyright infringement: an exploratory analysis

Abstract

Despite serious debate on Generative Artificial Intelligence (GenAI) and copyright infringement worldwide, between government regulation and organizations (artificial intelligence [AI] and creative organizations). There is a lack of in-depth understanding of users’ perspectives on GenAI and copyright infringement, which influences public opinion as well as government and organizational policies. This study aims to capture users’ perceptions of GenAI and copyright infringement as key factors driving policies and organizational products, using a mixed-methods research approach that incorporates qualitative (topic modeling) and quantitative (sentiment and statistical analyses). The study investigates copyright restrictions, training AI models with copyrighted data, ownership of AI-generated work and fair use. The researchers collected, extracted and cleaned 1,505 comments from Reddit using Python and the Reddit API and analyzed 1,435 comments. The researchers applied Qualitative (topic modeling) and quantitative (sentiment and statistical analyses) approaches to understand users’ perceptions of GenAI and copyright infringement. The results indicate that most users have positive sentiments, believe that GenAI infringes on copyrighted works, and that AI organizations or developers may depend on copyrighted works. Also, AI models are trained on copyrighted works. However, users suggest that there should be regulation and control of copyrighted work rather than a complete ban. The users show that there is no fair use of copyrighted works and suggest that the original holders of similar AI-generated output should be compensated. This study addresses copyright infringement and GenAI from the users’ perspective. Previous studies have not considered public opinion regarding GenAI and copyright infringement, which can shape policy and products.