Generative artificial intelligence now produces texts, images, and music that circulate as culturally meaningful artefacts, unsettling the evaluative vocabularies through which the social sciences and humanities describe authorship, creativity, and skill. Copyright law experiences this disruption as a problem of fit: its originality tests are built to recognise human intellectual creation, yet generative systems can imitate the surface features of human expression at scale. This paper asks whether the legal boundary between protected “original expression” and unprotected “algorithmic imitation” can be drawn coherently when the dominant indicators of creativity are increasingly output facing. It argues that common legal intuitions, which treat the Turing Test style indistinguishability of outputs as a proxy for creative intelligence, risk mistaking persuasive simulation for the kind of agency that originality doctrine presupposes. At the same time, relying on a simple human centred dismissal of machine production reproduces a comforting “heads in the sand” posture that obscures how contemporary models emulate aspects of human cognition through neural architectures and controlled randomness, and how institutions already treat their outputs as cultural and economic goods. To clarify this tension, the paper revisits two canonical thought experiments that continue to shape contemporary debates about artificial creativity: the Turing Test, which frames intelligence as behavioural success in interaction, and Searle’s Chinese Room, which separates symbol manipulation from understanding. Read together, they reveal an epistemic glitch in current regulatory practice: law asks for “intellectual creation” while lacking stable criteria for identifying, evidencing, or even conceptually locating it in hybrid human machine workflows. The paper further develops this glitch by drawing on Weizenbaum’s claim that computation is substrate independent, showing how attempts to ground creativity in biological embodiment can slide into expansive metaphysical positions that are unusable for legal governance. Building on this analysis, the paper proposes an interdisciplinary reframing for AIGC regulation, shifting from output based resemblance tests toward a modest set of epistemic and institutional indicators that law can operationalise in creative markets and cultural production. The aim is not to anthropomorphise machines, but to make the criteria for attributing creative value, responsibility, and entitlement more transparent in an age of automated cultural generation.

AI-Generated Creativity, Turing Test, Chinese Room Argument, Copyright Originality, Epistemic Fairness, Digital Humanities, Generative Artificial Intelligence, Authorship and Accountability