The impact of AI adoption on R&D productivity: Evidence from Chinese pharmaceutical manufacturing industry

Abstract

The decline in R&D productivity is a persistent and widespread phenomenon hindering long-term economic growth. With the boom of artificial intelligence (AI), especially the pervasive application of deep learning, AI is now able to participate in innovation, which is expected to help reverse the decreasing trend of R&D productivity. Focusing on the pharmaceutical manufacturing industry, one of the most active domains in AI-driven innovation, we take Chinese listed firms as an example to investigate the role of AI in drug discovery and the impact of AI adoption on new drug R&D productivity. Our empirical results show that other things being equal, new drug output per billion yuan invested in R&D averagely rises by 0.05–0.06, for each 1-unit increase in our AI-adoption index capturing the firm-level AI usage intensity. One of the mechanisms behind it, which we call “R&D elitism”, positively associates new drug R&D productivity with AI adoption through raising the share of core researchers (by 0.2 % on average in terms of educated staffs or 0.8 % on average in terms of experienced ones under the same condition as mentioned above) in the R&D team. With the power of AI, marginal researchers are getting harder to retain and new drugs are hence getting easier to find.