Nikolaus Thumm (Organisation for Economic Co-operation and Development; Technische Universität Berlin) & Gabriel Glänte (UC Berkeley Haas School of Business; AI Sweden) have posted AI, Innovation and the Future of the Patent System on SSRN. Here is the abstract:
Artificial intelligence is reshaping how inventions are created and how the value of innovation is appropriated, raising fundamental questions for the patent system. Drawing on twenty-seven semistructured interviews with practitioners operating at the frontier of AI-intensive research and development, intellectual property strategy, and patent practice across the United States and Europe, this paper examines two analytically distinct questions: how patents affect AI innovation, and how AI affects the patent system. It finds that the most commercially valuable AI assets, including training data, model weights, and engineering pipelines, are typically protected through trade secrecy rather than patents, with firms combining protection mechanisms across the layers of the technology stack. At the same time, AI tools are transforming patent workflows, including translation, prior art search, drafting, and examination. AI is also placing pressure on core patent doctrines, including the person-skilled-in-the-art standard, inventive step, disclosure, and inventorship. The paper introduces two conceptual frameworks—the Patentability-Value Gap and the Volume-Cost Scenario Matrix—as tools for structuring future research and policy discussion, and embeds comparative observations on the USPTO and EPO throughout. The analysis is exploratory and is offered as a structured snapshot of expert perspectives during a formative period, with each chapter identifying open questions for further empirical, doctrinal, and policy research.
Recommended!
To receive new posts from Legal Theory Blog by email, get a free subscription to Legal Theory Stack.
Lawrence Solum