Tejas Narechania, Professor

Closed (1) Machine Learning and Regulation

Applications for Spring 2019 are now closed for this project.

First among the new data-driven technologies that have important implications for personal privacy is machine learning (and artificial intelligence). But regulators know only very little about how the technology works, and about how it will evolve. As a result, the regulatory environment for such privacy-implicating technologies remains in flux and unstable. This project will unpack the underlying technology, by putting it in terms that are familiar to regulators and legal audiences.

The URA will help explain, in simple terms, how machine learning technology works. In particular, we'll focus on metrics such as cost, complexity, error, and rate of convergence, among others, in order to describe how machine learning compares to other sorts of computer algorithms --- and how it compares, on these dimensions, to traditional regulated industries (such as networked communications). This will require the URA to (1) become familiar with the basic of regulatory economics, (2) understand machine learning technology, and (3) identify and explain sources that describe the technology along the dimensions that are salient to regulatory theories.

Qualifications: An ideal URA will have excellent writing skills, a good working knowledge of computer science and machine learning algorithms, and an interest in law, policy, or economics.

Weekly Hours: 3-5 hrs