ÖйúÒ»¼¶Æ¬

We analyze the drivers and valuation of artificial intelligence (AI) innovations using a new dataset of AI patents developed by the USPTO. As the latest cluster of general purpose technology (GPT), we predict and find that AI patents are broadly developed across multiple sectors, but are increasingly concentrated in a subset of firms with a growing stock of knowledge in AI. On the valuation of AI patents, we find that AI patents tend to exhibit a 10% value premium compared to other patents, and this result is stronger in the IT industry and industries that are more suited for AI technologies. Based on Bresnahan (2010) definition of GPTs, we categorize and examine the valuation of the key types of AI innovations. We find that exploratory AI patents tend to be more highly valued compared to exploitative AI patents. Moreover, we find that general (application-specific) AI patents in IT (non-IT) industries tend to be more highly valued. Lastly, we find that AI patents are also associated with higher citations and better financial performance, suggesting that the potential for follow-on innovation, and new products and services could explain the higher valuations in AI patents.

zcarol2@ceibs.edu