AI Inventorship in Patent Law: Comparative Doctrine, Regulatory Divergence, and Harmonized Reform

Abstract

Artificial intelligence (AI) is increasingly embedded in invention workflows, yet patent systems continue to define inventorship in human terms. The central problem is not whether AI contributes to technical creativity, but whether existing legal doctrines can allocate ownership, responsibility, and evidentiary accountability when inventive processes are distributed across humans and AI systems. Prior scholarship has examined DABUS disputes and personhood debates, but the literature remains fragmented across jurisdictions and often under-specifies operational reform pathways for patent offices and courts. This article addresses that gap through a comparative doctrinal design covering the United States, United Kingdom, European Union/Germany, Australia, South Africa, India, China, and selected emerging jurisdictions. The analysis triangulates statutes, judicial decisions, patent-office guidance, and Scopus-indexed legal scholarship from 2019–2026. The findings show convergence on human-only inventorship, but persistent divergence in disclosure expectations, contribution tests, and evidentiary practices. These divergences produce transparency deficits, cross-border uncertainty, and uneven enforcement risk for multinational portfolios. The article’s original contribution is a six-part harmonized reform model that preserves human-centered inventorship while modernizing AI-contribution disclosure, contribution assessment, evidentiary protocols, interoperability standards, targeted supplementary protection for edge cases, and governance integration. The model is explicitly grounded in an accountability architecture framework that separates legal status from evidentiary governance. This framework offers a practical route to doctrinal stability and policy coherence without requiring full AI legal personhood.