The opacity, unpredictability, and inscrutability of recent computer systems continues to challenge users and regulators. While one approach has been to work toward explainability and transparency, this tactic is expensive and has not yet proven to be broadly successful. This talk proposes that recent developments in artificial intelligence and machine learning might instead be usefully understood as a new “style” of knowledge in the manner used by historian Ian Hacking (1990) to explain the role of technology in the history of science. Rather than containing a nugget of information meant to be revealed from inside a black box, in this way of thinking these systems might be a new form of argumentation and delegation in themselves. This suggests a different emphasis for future policymaking.

Bio:

Christian Sandvig is a visiting professor at the School of Social Science of the Institute for Advanced Study (IAS) this academic year. While at IAS, he will work on a book project about the ethics and consequences of algorithmic filtering and machine learning used within online platforms.

Sandvig is also the director of the Center for Ethics, Society, and Computing (ESC) and collegiate professor of Information at the University of Michigan. His research about civil rights and discrimination online has been cited by the US Supreme Court (2021) and the Obama White House (2016). He has worked in the private sector as a software developer in addition to his experience in academia.

This talk was not recorded.

If you need an accommodation for a disability please contact Jean Butcher at at least one week before the event.