Politics, Ethics and AI Technology: The State of the Discipline

Session Submission Type: Full Paper Panel

Session Description

Artificial intelligence (AI) is supposed to be changing the world, but it is not always clearly for the better. In recent years, this suite of technologies has been accused of stealing jobs and driving economic discontent, enabling foreign powers to meddle in elections, transforming the nature of democratic elections, and reinforcing discrimination in education and policing. Scholarly approaches to studying how AI is transforming our social and political world have mostly been confined to the fields of practical ethics, law and computer science. Political scientists and political theorists have not yet developed systematic approaches to exploring and evaluating how politics shapes AI, and how AI shapes politics.

This panel will explore how political scientists and political theorists should study information technology – and particularly AI. It will examine in a systematic fashion, four of the central debates that are currently driving research in AI in law, computer science and ethics. The first considers how an institution’s policy goals, in this case in policing, shape the way artificial intelligence is designed and used. The second explores theories of justice when artificial intelligence is used to make decisions. The third concerns how institutions should explain and justify choices made using artificial intelligence. The fourth points to the limits of studying these all of these issues using an ‘AI ethics’ framework, arguing for a focus on political theory and political economy instead.

The panel gathers leading scholars in political theory, philosophy, computer science and law to present their papers. Its goal goes beyond discussing papers, however. The aim of this panel is to begin to build a field, a systematic scholarly approach to studying how political and economic institutions are shaping the development of technology, and how, in turn, the development of technology is shaping political and economic institutions. It is politics and political economy perspectives that the current study of AI sorely lacks.

This subject should be of interest to a broad range of scholars at the American Political Science Association: those interested in the nature and scope of technological change, particularly related to AI; those who want to understand how other disciplines are studying these phenomena; and particularly, those who feel that these current approaches need to better take account of existing literatures in political science and political theory.

Sub Unit

• DIVISION 2: Foundations of Political Theory

Individual Presentations

• The Paradoxes of Predictive Policing - Solon Barocas, Cornell University

• Human Judgement in Algorithmic Loops - Reuben Binns, University of Oxford

• A Philosophy for the Science of Explainable AI - Rune Nyrup, University of Cambridge

• A Political Theory Approach to ‘AI Ethics’ - Joshua Simons, Harvard University; Chloé Bakalar, Temple University

Chair

• Chloé Bakalar Temple University

Discussant

• Erin Miller Princeton University