Digital Sweat from the Virtual Brow: an Examination of Generative AI Labor and Production of Intellectual Property

Abstract

Generative artificial intelligence (genAI) tests the limits of copyright law in the United States (US), requiring the US Copyright Office to reassess its longstanding position on computer-made works. The Copyright Office argues against awarding copyright to genAI works because it overextends the limits of Sweat of the Brow Doctrine, which rewards effort with ownership and protection because the result has value. Sweat of the Brow Doctrine is based on John Locke’s labor theory of property, which posits that mixing labor with resources yields property and rights to said property; however, awarding copyright to genAI works would deplete the commons due to the incredibly high rate of productivity exhibited by genAI. This article reinterprets Locke’s labor theory of property and proposes an extension to the framework. This extended framework is then applied to a hypothetical scenario that shows how technology firms may use genAI, streaming media platforms, and automated copyright enforcement systems to flood websites with genAI content, enabling them to assert claims over similar human-authored works. The result of this scenario is expanded to the larger terrain of US copyright law, showing that such privatization of content would impoverish the public domain and exploit unpaid human creativity. Under the proposed extended theory, AI outputs belong in the public domain, demonstrating how the Copyright Office’s argument ultimately protects the commons.