Prithwiraj Choudhury, Harvard Business School
Prithwiraj Choudhury, Harvard Business School
TOM Seminar Experimental Evidence on Productivity Complementarities between Human Capital and Machine Learning
TOM Seminar Experimental Evidence on Productivity Complementarities between Human Capital and Machine Learning
29 Mar 201812:30 PM – 2:00 PM
Faculty and doctoral students only
The advent of artificial intelligence in the form of
machine learning technologies raises questions regarding productivity of
workers adopting such technology and potential complementarity with prior human
capital. It is plausible that the adoption of machine learning in predictive
tasks raises the value of certain skills (such as computer science) and lowers
the value of certain skills (such as reading). Within the research context of
patent examination by the US Patent and Trademark Office, which has developed
both machine learning and Boolean search process technologies to predict
relevant prior art, we randomly assign both technology and expert advice
constituting how to use technology and task specific information. We track
patent examination outcomes in which the goal is to identify prior art that
invalidates claims being examined. We find heterogeneous productivity effects
for workers with and without computer science and engineering (CS&E)
backgrounds. Conditional on receiving expert advice, participants without
CS&E backgrounds are less productive with the machine learning technology,
while those with CS&E backgrounds are more productive with it. We examine
underlying mechanisms and present evidence suggestive that participants without
a CS&E background benefit from using older technology due to their skills
in language comprehension. We also present evidence that the higher
productivity of CS&E workers using the machine learning technology is
related to their skills in manipulating software interfaces. We conclude that
the adoption of this machine learning process technology lowers the relative
value of skills such as reading in predictive tasks and raises the relative
value of CS&E skills. These differentials are exacerbated and not mitigated
by the provision of expert advice.
Location:
Cotting Conference Room
Organizer:
Prithwiraj (Raj) Choudhury is an Assistant Professor in the Technology and Operations Management Unit at the Harvard Business School. He was an Assistant Professor at Wharton prior to joining HBS. He studies knowledge worker productivity and innovation, with a focus on studying how global R&D, talent flows and AI will shape the future of work.