Fair-Use Doctrine: Copyright Challenges Posed by AI Generative Technology and In-Text and Data Mining Training

Abstract

Artificial intelligence (AI) innovators have invested billions in research and development to create advanced software and hardware tools. AI has generated new businesses and start-ups, providing employment to millions. However, despite its transformative potential, AI innovation has not received sufficient support from legal systems and remains constrained by the current intellectual property regime. Fair-use exemptions, particularly with respect to copyright law, have posed challenges for AI development and training. This disconnect arises because copyright law has not evolved in tandem with AI technologies and still reflects principles from an earlier computing era. Consequently, there is a pressing need to examine which provisions of existing copyright frameworks may be impeding AI progress, especially those related to fair-use exceptions. The ambiguity surrounding these exceptions has led to unpredictable judicial interpretations, particularly in the context of AI tools and technologies. As numerous generative AI systems, including OpenAI’s ChatGPT, rely on large datasets that incorporate both copyrighted and non-copyrighted materials, the process of AI training has become a focal point of legal, ethical, and artistic debate. This paper explores these complexities and examines the emerging copyright challenges associated with the training and use of generative AI systems.