The Dream of the Race-Blind Algorithm

Redistricting -- drawing boundaries for Congressional and other electoral districts -- has long been an area where the U.S. Supreme Court has worked to develop a jurisprudence that centers "race-blind" comparisons.  Alongside employment and college admissions, redistricting has produced complicated and consequential decisions that hinge on who can take what information "into account" in their work.  Since at least 1982, authors thinking about how to use the Voting Rights Act have wondered if computers can deliver the right kind of blindness.  I'll tell the story of the last decade of Supreme Court struggles with the promise and perils of algorithms in redistricting.  

Moon Duchin is a professor of Math and Public Policy at Cornell, and she runs the Data and Democracy Lab in the Brooks School.  She's worked as an expert witness in numerous redistricting cases in the last census cycle.  She loves to think about how ideas and artifacts from math and computing operate persuasively in public opinion and in courts of law.  Duchin was one of the founders of the STS major at Tufts before moving to Cornell.