Expression, Verification, and the Scarcity of Aesthetic Education: Reinterpreting Moral Rights in the Age of Generative AI

Abstract

This paper argues that the denial of copyright and moral rights to works substantially shaped by human expressive judgment in generative AI systems produces two interrelated harms: the incentive to conceal AI involvement (“trace forgery”) and the proliferation of low-quality content (“flooding”). Drawing on R. G. Collingwood’s distinction between craft and expression in The Principles of Art, it contends that generative AI separates technical execution (craft) from the human process of clarifying an initially indeterminate intention (expression). Chinese judicial practice, which requires process records of prompting, iteration, selection, and modification, already partially recognizes this separation. In the AI era, the functions of moral rights shift: the right of attribution becomes a declaration of authorial acceptance or verification, while the right of integrity protects the author’s evaluative standard and associated reputation. Recognition of these functions aligns incentives toward the production of excellent works and open acknowledgment of AI participation, rather than concealment or flooding. The stakes are uniquely high because the present generation of children is forming aesthetic judgment precisely when craft and expression have been decoupled; the window for correct training of expressive capacity is therefore the true cost of inaction. Historically, copyright has regulated successive scarcities (books, authors, channels, attention), yet its underlying object has always been educational opportunity. The current scarcity is that of proper aesthetic education itself.