Eugenie Dugoua
London School of Economics (Visiting UC Berkeley ARE)
Monday, October 20, 2025
4:10-5:10pm
241 Giannini Hall, UC Berkeley
“Directed Technological Change and General Purpose Technologies: Can AI Accelerate the Clean Energy Innovation?”
Abstract: Accelerating the rate of clean innovation relative to dirty innovation is critical to decarbonization. This race between clean and dirty technologies is taking place against the backdrop of improvements in general-purpose technologies (GPT) such as information and communication technologies (ICT) and artificial intelligence (AI). Using patent data, we find that clean energy technologies enjoy more spillovers from AI and ICT than dirty energy technologies, and that these increase the value of patents. Theoretically, we show that this differential rate of spillovers from a GPT should induce firms to redirect their efforts towards clean technologies. This would suggest that exposure to AI/ICT may be a new channel encouraging clean innovation. We then explore this causal mechanism using a shift-share instrument exploiting changes in firms’ exposure to AI R&D based on the prior geographical location of their energy inventors. We discuss policy implications using a directed technical change model: we show that this induced clean innovation channel can reduce the carbon tax needed to achieve emission targets, but that AI R&D subsidies could crowd out clean R&D efforts without commensurate increases in clean R&D subsidies.