What influences attitudes about artificial intelligence adoption: Evidence from U.S. local officials

Abstract

Rapid advances in machine learning and related techniques have increased optimism about self-driving cars, autonomous surgery, and other uses of artificial intelligence (AI). But adoption of these technologies is not simply a matter of breakthroughs in the design and training of algorithms. Regulators around the world will have to make a litany of choices about law and policy surrounding AI. To advance knowledge of how they will make these choices, we draw on a unique survey pool—690 local officials in the United States—a representative sample of U.S. local officials. These officials will make many of the decisions about AI adoption, from government use to regulation, given the decentralized structure of the United States. The results show larger levels of support for autonomous vehicles than autonomous surgery. Moreover, those that used ridesharing apps prior to the COVID-19 pandemic are significantly more supportive of autonomous vehicles. We also find that self-reported familiarity with AI is correlated with increased approval of AI uses in a variety of areas, including facial recognition, natural disaster impact planning, and even military surveillance. Related, those who expressed greater opposition to AI adoption also appear more concerned about trade-offs between privacy and information and bias in algorithms. Finally, the explanatory logic used by respondents varies based on gender and prior experience with AI, which we demonstrate with quantitative text analysis.