Artificial Intel: Pol Prob

3 Credits · Level: Graduate, Undergraduate

This class considers the policy implications of artificial intelligence (AI). It teaches students the basics of how artificial intelligence algorithms work and how they can go awry, particularly in real world applications. No prior technical knowledge of artificial intelligence is needed to be successful in the class. The class considers the interaction between technical experts and subject matter experts as AI is implemented in a specific domain: how much friction will accompany that implementation? Will the unique challenges of implementing AI in policing or autonomous cars or the life sciences weaken or even undercut the general analytical power of marquee AI algorithms? What types of errors occur and what are their consequences and what are strategies to measure and mitigate them in different domains? The class considers broader policy questions including whether AI inherently favors autocracies by closing the space for political dissent through persistent surveillance and the greater ease with which they can gather and use large databases of various sorts of information. The class studies various policy frameworks, some implemented and most merely posited, to maximize benefits from the use of these algorithms and to minimize their risks. The implemented frameworks include, among others, the frameworks to regulate AI in countries including the United States, China, and the European Union.