When do Citizens Resist AI-Usage in Public Policy? Evidence from the COVID-19 Crisis

Abstract

In the effort to contain the COVID-19 outbreak and mitigate its consequences, many hopes were placed on the use of AI-technology. Indeed, AI-based algorithms have informed decision-making in many aspects of the crisis, including diagnosis, infection tracking, outbreak forecasting, and allocation of essential resources and services. How do citizens view the incorporation of AI technology in making major public policy decisions? Using original survey data from the U.S. and Israel, I show that the public trusts humans significantly more than algorithms in making pandemic-related highstake decisions. However, experimental evidence indicates that this clear preference does not translate uniformly into lower support for policies that rely on algorithmic assessment. Rather, views on the use of algorithms differ substantially depending on the decision context in which it is deployed. The analysis indicates that while deployment of AI in public policy may be acceptable in some cases, it could trigger substantial opposition in other, theoretically-predictable contexts.

Keywords: algorithmic decision-making, public policy, COVID-19, public attitudes, algorithmic fairness