Patents are often seen as indicators of innovation, and to foster sustainable innovation, it is essential to gain deeper insights into patents’ environmental impact. They can be modelled to estimate their carbon footprint when implemented, but this process is often time-consuming and resource-intensive. This article explores a new method for assessing the carbon footprint of patents using large language models (LLMs). In this approach, patent data is analysed to extract relevant keywords, which are then linked to their associated carbon emissions using LLMs. Once the carbon estimates for all patents in a portfolio are obtained, one can rank them based on their carbon footprint.

1 . Climate Change

Climate change is an undeniable reality, placing humanity at a critical juncture. The decisions we make in this century will shape the future of our planet. While governments, corporations, and institutions are actively pursuing net-zero emissions and carbon neutrality, these efforts alone may not be enough. To make a lasting impact, robust policies combined with cutting-edge technological solutions that address both climate mitigation and adaptation should be adopted. In his influential work, Michael Grubb1 argues that a comprehensive understanding of the economics of technological innovation offers a way forward between what seem very divergent international positions on climate change policy. The United Nation’s adoption of the 17 sustainable development goals (SDGs) laid a global framework for tackling climate change. Of specific interest when it comes to innovation is SDG Goal 9 which focuses on “industry, innovation, and infrastructure”. The aim of SDG Goal 9 is to develop resilient infrastructure, encourage sustainable industrialization, and foster innovation. Patents are often seen as indicators of innovation, and to foster sustainable innovation, it is essential to gain deeper insights into patents’ environmental impact.

2 . Green Patents

Patents related to environmentally friendly technologies, or “green patents,” are vital in the fight against climate change. The World Intellectual Property Organization (WIPO) GREEN online platform facilitates technology exchange by linking providers and seekers of environmentally friendly technologies. Green patents are classified using frameworks such as the ENV-TECH by the OECD, the IPC Green Inventory by WIPO, and the Y02/Y04S scheme by the European Patent Office (EPO). These classification systems categorize patents based on environmentally friendly technologies related to energy, greenhouse gas (GHG) capture, transport, and waste management. Lauren Cohen et al.2 document a notable trend in green patent production, finding that firms in the oil, gas, and energy sectors — often with lower Environmental, Social, and Governance (ESG) scores and frequently excluded from ESG investment funds — are leading innovators in the U.S. green patent landscape.

3 . Patent Ranking Based on Carbon Footprint

Traditionally, patents are ranked based on metrics like citations, patent utilization, market coverage, and technology relevance. Sooyoung Oh et al.3 propose evaluating and ranking patents using weighted citations. Another notable work is by Ernst and Omland4 that discusses creating a patent asset index as a key metric to evaluate the value of a firm’s patent portfolio by combining three factors: portfolio size, market coverage, and technology relevance.

A recent trend is to evaluate and rank patents based on their carbon footprint. The carbon footprint of a patent refers to the estimated carbon emissions generated by the systems, processes, or products described in the patent. If a patent is reduced to practice, its carbon footprint reflects the actual emissions released during its use or production. New tools, such as a patent carbon footprint calculation tool5, allow for quick analysis of patent text to identify keywords linked to carbon emissions. These keywords, generated by large language models (LLMs), are mapped to corresponding emissions data. If necessary, users can manually input additional keywords and the tool will calculate the carbon emission automatically. The carbon emission data usually is activity-based emission data or spend-based emission data. If the data is spend-based emission data, then the corresponding carbon emission for the spending activity is calculated, as illustrated in Fig. 1. Essentially, using tools5 each patent can be assigned an estimated carbon emission value, which in turn helps in ranking the patents, as shown in Table 1.

4 . Conclusion

The carbon impact of patents offers a new dimension for assessing their value. By determining the carbon emissions of patents, companies can make informed decisions during patent drafting, rank patents, enhance their ESG reporting, and improve the sustainability of their patent portfolios. As the world pushes for more sustainable innovation, understanding the carbon footprint of patents will play a crucial role in shaping the future of green technologies.

Fig. 1. Sample keywords generated and the corresponding carbon emission data extracted by the LLM

 

Emission Activity name
(Suggested Keywords)		 Carbon Emission
(Co2e in Kg)		 Patent Ranking
Divide electric devices into power groups;  store correction amounts in the controller; perform increase process of engine rotation speed 15 1
Control of a vehicle; relay data; detect vehicle state;
specify controller for low transmission priority;
suppress communication of data		 326 2
Warming vehicle components; distribute fluid through coolant path; heating engine, transmission, and battery 821 3
Fuel cell system includes fuel cell, high voltage circuit and relay; driving an electromotor with high voltage circuit; electrically connecting or blocking fuel cell to high voltage circuit; performing conductivity reduction process on cooling liquid 2121 4
Maintain SOC level of the battery; control an engine, a motor, a steering system, a brake system; 2723 5

Table 1: Patent Ranking based on Carbon Emission

Bibliography

1 . Grubb, M. (2004). Technology Innovation and Climate Change Policy: an overview of issues and options. Keio Economic Studies, 41(2), 103-132.

2 . Cohen, L., Gurun, U. G., & Nguyen, Q. H. (2020 (Revised 2024)). The ESG-Innovation Disconnect: Evidence from Green Patenting. NBER Working Paper Series.

3 . Oh, S., Lei, Z., Mitra, P., & Yen, J. (2012). Evaluating and ranking patents using weighted citations. JCDL ’12: Proceedings of the 12th ACM/IEEE-CS joint conference on Digital Libraries. ACM Digital Library.

4 . Ernst, H., & Omland, N. (2011, March). The Patent Asset Index – A new approach to benchmark patent portfolios. World Patent Information, 33(1), 34-41.

5 . (2024, May). ESGeniX. Retrieved from ESGeniX: https://esgenix.com

 About the Author

Chemmannore Renooj Jacob is an intellectual property professional with more than 20 years of experience. Mr. Jacob founded a boutique Intellectual Property firm, IP Astra, which was later merged with a top-tier intellectual property law firm, LexOrbis. Mr. Jacob was the Director at BPE Biotree India Pvt Ltd, a biopharma equipment manufacturer, now a member of Zeta Group, Austria. His latest venture aims to bring a paradigm shift in how patents are seen from a sustainability point of view. Mr. Jacob has a B.E, and M.Tech degree in electronics and has done an executive program in global business management.