Many of copyright’s core concepts—from authorship and ownership to infringement and fair use—are being challenged by the rapid rise of generative AI. Whether in service of creativity or capital, however, copyright law is perfectly capable of absorbing this latest innovation. More interesting than the doctrinal debates that AI provokes, then, is the opportunity it presents to revisit the purposes of the copyright system in the age of AI. After introducing the AI-copyright challenge in Part 1, Part 2 considers the guiding principles and normative objectives that underlie—and so ought to inform—copyright law and its response to AI technologies. It proposes a substantive approach to tech-neutrality aimed at achieving normative equilibrium in the face of technological disruption. Applying this frame—with its corresponding emphasis on authorship and the public interest—Part 3 goes on to explain why AI-generated outputs are therefore uncopyrightable and AI-training inputs are non-infringing.