This project investigates the impact that AI has on the pace of green innovation and that ability of clean technologies to catch up faster with dirty ones.
Currently, we use patent data to quantify spillovers from AI to green technologies. Acemoglu et al. (2016) as well as this recent paper by Pichler et al. show that spillovers as measured by patent citations are surprisingly good predictors of the rate of innovation in a particular field. So, patents are a good place to start.
We show that AI, as a general-purpose-technology, can significantly affect the dynamics of the green transition. We find that green technologies are better able to make use of AI than dirty ones, which is great news.
We are furthering the analysis to understand how policies that increase the rate of application of AI to green technologies might affect overall carbon emissions.
We are also investigating who creates AI inventions that are useful to green tech and who is able to apply them to green tech: do the spillovers happen primarily within firms and if so, what type of firms? or do they occur across firms but within regional networks of innovation? do large tech companies play an important role?
In collaboration with Pia Andres and Eugenie Dugoua.