Recalibrating Intellectual Property Protection in the Age of Generative Artificial Intelligence A Case Study of the Novara–Aurelis Dispute

Abstract

The rapid development of generative artificial intelligence has created significant uncertainty regarding the appropriate boundaries of intellectual property protection. Conventional copyright and patent frameworks were largely designed around human creators, identifiable inventors, relatively stable production processes, and distinguishable relationships between authorship and ownership. Generative artificial intelligence systems challenge each of these assumptions. This article examines these challenges through a fictional case study concerning the dispute between Novara Labs, an artificial intelligence company, and Aurelis Media Group, a multinational digital content enterprise. The dispute concerns the alleged unauthorized use of copyrighted training materials, ownership of machine-generated creative outputs, and the attribution of intellectual property rights to technological systems that operate with varying degrees of human intervention.
The study adopts a qualitative case-study methodology based on documentary analysis. The hypothetical dispute is reconstructed through contracts, licensing records, internal governance documents, correspondence, technical reports, and judicial submissions. The analysis identifies three principal dimensions of contemporary intellectual property uncertainty: the legality of training artificial intelligence models on protected works, the allocation of rights in outputs generated through artificial intelligence systems, and the evidentiary difficulties associated with establishing substantial similarity and causal relationships between protected works and generated outputs. The case further demonstrates that conventional distinctions between input and output infringement may be insufficient when artificial intelligence systems operate through complex and partially opaque computational processes.
The article argues that intellectual property law should not respond to generative artificial intelligence through a single, technology-specific rule. Instead, a layered regulatory approach is required, combining stronger transparency obligations, contractual licensing mechanisms, human-centred authorship standards, collective licensing structures, and evidentiary reforms. The case study contributes to the emerging literature by illustrating how these issues interact within a single commercial dispute and by identifying practical governance mechanisms that could reduce uncertainty for both rights holders and technology developers.