Abstract
The ever increasing use of generative artificial intelligence raises significant questions about authorship and related issues such as credit and accountability. In this paper, I consider whether works produced by means of users inputting natural language prompts into Generative Adversarial Networks are works of authorship. I argue that they are not. This is not due to concerns about randomness or machine-assistance compromising human labor or intellectual vision, but instead due to the syntactical and compositional limitations of existing AI systems in handling natural language prompts. This, I argue, gives rise to ‘authorship gaps’.
Generative Adversarial Networks (GANs), authorship, intellectual property rights, syntax, compositionality
© 2024 by the Board of Trustees of the University of Illinois
2024
You do not currently have access to this content.