Artificial Intelligence Policy Challenges

Abstract

Artificial intelligence and machine learning are becoming increasingly integrated into our economy and society. This integration creates a range of challenges policymakers will have to address. Probably the most-discussed question is the effect AI might have on the composition of the labor market and how we should react. Some advocate slowing or taxing shifts to automation, but most accept the productivity-enhancing aspect of these advances and worry instead about how to compensate workers who end up on the losing side of the changes.

But the future of work is not the only relevant policy issue. For example, in a twist on the jobs question, AI and machine learning create new ways to analyze data on labor markets. More broadly, how should government incorporate these new tools into its own analytical capabilities? Who should be liable in the event of accidents involving autonomous vehicles? This panel will discuss these and other issues facing policymakers as the technology advances.

This panel is likely to be relevant to the TPRC community for a number of reasons. First, the topic is a new technological development with major policy implications. Second, it is related to other areas covered by the conference: privacy, inequality related to digitization, intellectual property (related to questions of algorithmic transparency), and data science.

The panel will be comprised of scholars who participated in a recent event sponsored by the Technology Policy Institute, which featured nine new papers. The speakers cover different, but related, topics:

• Anton Korinek, University of Virginia: Research, with Joe Stiglitz, focuses on the potential implications of AI on inequality. • Maja Brkan, Maastricht University: Privacy laws and algorithms. • Jeff Prince, Indiana University: Assessing the scope and performance of predictive analytics. • Sarah Oh, Technology Policy Institute: Machine learning and the regulatory process. • Scott Wallsten, Technology Policy Institute: Moderator.

We expect this to be a lively panel that addresses the policy issue that occupies most of the discussion (future of work) but expands the discussion to encourage people to think more deeply about the range of policy issues these new technologies generate.