Authorship and Patentability of AI Generated Inventions: Rethinking the Human Requirement in IP Law

Artificial intelligence now materially contributes to inventive activity in fields such as drug discovery, advanced materials, and control systems, yet most patent regimes still tie inventorship to natural personhood, creating friction between doctrine and practice. This paper asks whether and how patent law should accommodate AI-generated and AI-assisted inventions without extending legal personality to machines. Using a doctrinal and comparative approach across the United Kingdom, United States, European Union, and India, it examines recent jurisprudence and administrative practice that reject naming AI as an inventor while permitting patents where a human meets the conception threshold and demonstrably integrates AI outputs into the claimed solution. The analysis situates national trends within ongoing multilateral discussions on disclosure, enablement, and entitlement, focusing on mechanisms that improve public notice and reproducibility when AI plays a substantive role. Building on accountability and incentive rationales, the paper proposes a hybrid inventorship framework that preserves human inventorship, mandates calibrated disclosure of AI’s contributions across problem framing, output selection, and validation, and allocates default rights to the human organizer or controller of the AI-enabled inventive process. This approach aims to deter under‑disclosure, reduce forum shopping, and stabilize ownership chains, while maintaining administrability and respecting legitimate confidentiality for proprietary models beyond what enablement requires. The conclusion contends that harmonized soft‑law tools and model provisions can deliver near‑term convergence under existing treaties, preparing the ground for targeted statutory refinements as AI capabilities and industry practice evolve.