A Tiered Framework for Copyright Ownership in AI-Generated Content

Abstract

The rapid proliferation of Artificial Intelligence Generated Content (AIGC) has precipitated an ontological crisis in copyright law, challenging its foundational anthropocentric principles. This study constructs a tripartite subjecthood framework to resolve attribution dilemmas across the autonomy spectrum of generative systems. Through doctrinal analysis of seminal cases (Naruto v. Slater, Feist v. Rural) and emergent legislation (EU AI Act, China’s Interim Measures), we demonstrate that granting AI legal personhood fundamentally conflicts with copyright’s utilitarian purpose. Crucially, we introduce a ​​Contribution Weight Matrix​​ (α·HI + β·AD + γ·UP) quantifying human-algorithmic collaboration in hybrid creation scenarios, validated against ISO/IEC 23053-2 documentation standards. Our legislative proposal advocates: Mandatory blockchain provenance registration(C2PA standard); Sui generis rights for transformative generativity; Developer strict liability-training data infringement. Empirical evidence from Vanity Fair v. AI Art Collective and Japan’s 2025 Copyright Act confirms the framework’s cross-jurisdictional viability. This research provides the first computational solution to AIGC copyright allocation while preserving creative incentives in the algorithmic age.