OECD Technical Paper Proposes New Patent-Based Method to Identify Emerging AI Technologies
SDG 9: Industry, Innovation and Infrastructure | SDG 17: Partnerships for the Goals
Institutions: Department of Science & Technology | Ministry of Electronics & Information Technology
The OECD has published a technical paper presenting a semi-automated methodology to identify AI patents, combining Cooperative Patent Classification (CPC) codes with an enriched list of AI-related keywords to capture emerging innovations; especially in generative AI and image processing.
Under the revised taxonomy, the number of AI patents identified increases by over 50% relative to the earlier OECD method, especially by including technologies previously excluded (e.g. those using new AI vocabulary). The methodology is intended to be adaptable and updatable (about bi-annual refresh) to keep pace with AI developments.
The report analyses trends in AI patenting (2010β2023), noting that in 2023, nearly 6% of PCT patent filings embedded AI components, a threefold increase over 2015. It also highlights concentration: China, the US, and Japan accounted for the bulk of AI patents in recent years.
This methodology offers Indiaβs technology agencies (DST, MeitY, DPIITβs innovation desks, Indian Patent Office) a tool to track emerging AI subfields more systematically, rather than relying on ad hoc signals. By identifying nascent AI areas (e.g. new generative architectures) early, India can target research funding, IP incentive schemes, cluster development, and regulatory foresight in those fields. From a governance perspective, having a clearer picture of AI patenting trajectories helps in anticipating technology governance needs (data, algorithm regulation, AI safety) in advance, rather than reactive policy.
Follow the full document here: OECD β Identifying Emerging AI Technologies Using Patent Data (PDF)