How AI Automation Affects Workers’ Performance and Political Preferences

• In Event: AI and the Future of Democracy

Abstract

Artificial intelligence (AI) is becoming an integral part of the workplace. AI-based applications are increasingly used to assist with or carry out important managerial tasks such as evaluating resumes, allocating tasks, and assessing productivity. How do workers respond to algorithmic management? How does this experience shape their attitudes on policies designed to deal with transformations in the labor market? To date, little is known about these questions. We conducted a field experiment in an online labor marketplace, where we randomly assigned 1500 workers to either human or algorithmic managers and varied also the type of interaction with those managers (either positive or negative). Our results indicate that working under algorithmic rather than human management has significant consequences on the way workers approach the job, a difference that varies depending on the type of interaction with the manager. Workers who receive positive feedback from human managers put significantly more effort and thought into their tasks, perform more accurately, and exhibit a higher commitment compared to workers who receive positive feedback from algorithmic management. Using data from a follow-up survey, we examine the implications of our findings for workers’ support for policy responses to AI automation, such as retraining programs and government compensation schemes.

Authors

• Shir Raviv, Columbia University

• Yotam Margalit, Tel Aviv University