Green intelligence: the AI content of green technologies

Abstract This paper investigates the contribution of artificial intelligence (AI) to environmental innovation. Leveraging a novel dataset of USPTO patent applications from 1980 to 2019, it explores the domain of green intelligence (GI), defined as the application of AI algorithms to green technologies. Our analyses reveal an expanding landscape where AI is indeed used as a general-purpose technology to address the challenge of sustainability and acts as a catalyst for green innovation. We highlight transportation, energy, and control methods as key applications of GI innovation. We then examine the impact of inventions by using measures and econometric tests suitable to establish (1) how AI and green inventions differ from other technologies and (2) what specifically distinguishes GI technologies in terms of quality and value. Results show that AI and green technologies have a greater impact on follow-on inventions and display greater originality and generality. GI inventions stand out even further in these dimensions. However, when we examine the market response to these inventions, we find positive results only for AI, indicating a mismatch between the technological vis-à-vis market potential of green and GI technologies, arguably due to greater uncertainty in their risk-return profiles.