Explainable Artificial Intelligence: Can We Hold Machines Accountable?

Tags: Technology Policy

View Event Recording.

With artificial intelligence (AI) becoming more ubiquitous, an increasing number of significant decisions are being made by machines. For example, some companies follow the recommendations of AI systems when hiring. In the legal world, judges increasingly rely upon AI systems in bail or sentencing decisions for criminal defendants. A job applicant or a criminal defendant might reasonably wonder: why did the AI system come to the decision that it did? The reality is that many of these AI decisions are difficult for humans to understand and AI-based outcomes cannot always be explained. Are there ways to make AI decisions more explainable, more understandable, and more accountable?

Legal scholars have characterized AI (machine-learning) decision-making as “black box” decision-making, and note that it raises problems of fairness, legitimacy, and error. While computer science has long discussed the concept of explainability, different notions—for example, making algorithmic decisions understandable to individuals subject to those decisions—have taken hold in the legal community. Even within recent legal scholarship, different concepts of accountability and explanability abound.

This conference will advance the state of knowledge surrounding AI and explainability for multiple constituencies, including private sector firms that are actively developing artificial intelligence systems, governments and policymakers who are navigating possible regulatory approaches in this area, academic entities who are studying this space, and the public at large.


Sessions

05/03/19 9:30am - 9:35am
Welcome
  • Harry Surden
    Interim Executive Director, Silicon Flatirons; Associate Professor of Law, University of Colorado Law School
05/03/19 9:35am - 10:40am
What Problem is AI Explainability Supposed to Solve?
  • Nicholson Price — Moderator
    Assistant Professor of Law, University of Michigan Law School
  • Mireille Hildebrandt — Panelist
    Research Professor, Law Science Technology & Society, Vrije Universiteit Brussel
  • Pauline Kim — Panelist
    Daniel Noyes Kirby Professor of Law, Washington University in St. Louis School of Law
  • Max van Drunen — Panelist
    Faculty of Law, University of Amsterdam
  • Frederik Zuiderveen Borgesius — Panelist
    Professor of Law, Digital Security Group, Institute for Computing and Information Sciences, Radboud University
05/03/19 10:40am - 11:00am
Break

05/03/19 11:00am - 12:10pm
Explainability in Computer Science versus the Law
  • Harry Surden — Moderator
    Interim Executive Director, Silicon Flatirons; Associate Professor of Law, University of Colorado Law School
  • Kiel Brennan-Marquez — Panelist
    Associate Professor of Law, William T. Golden Scholar, University of Connecticut School of Law
  • Bradley Hayes — Panelist
    Assistant Professor, Computer Science, University of Colorado Boulder
  • Joshua A. Kroll — Panelist
    Postdoctoral Scholar, School of Information, University of California at Berkeley
  • Pierre Schlag — Panelist
    University Distinguished Professor, Byron R. White Professor of Law, University of Colorado Law School
  • Mary-Anne Williams — Panelist
    Distinguished Professor, University of Technology Sydney; Director, The Magic Lab
05/03/19 12:10pm - 1:30pm
Lunch

05/03/19 1:30pm - 2:40pm
Comparative Legal Approaches to Explainability
  • Margot Kaminski — Moderator
    Associate Professor, University of Colorado Law School; Privacy Initiative Director, Silicon Flatirons
  • Meg Leta Jones — Panelist
    Associate Professor, Georgetown University
  • Gianclaudio Malgieri — Panelist
    PhD Researcher, Law Science Technology & Society, Vrije Universiteit Brussel
  • Andrew Selbst — Panelist
    Postdoctoral Scholar, Data & Society Research Institute
  • Carlos Affonso Souza — Panelist
    Director, Partner, Instituto de Tecnologia & Sociedade do Rio de Janeiro
  • Michael Veale — Panelist
    Technology Policy Researcher, University College London
05/03/19 2:40pm - 3:00pm
Break

05/03/19 3:00pm - 4:10pm
Explainability: A Way Forward?
  • Deven Desai — Moderator
    Associate Professor, Area Coordinator for Law and Ethics, Scheller College of Business, Georgia Institute of Technology
  • De'Aira Bryant — Panelist
    PhD Student, Computer Science, Georgia Institute of Technology
  • Brent Mittelstadt — Panelist
    Research Fellow, British Academy Postdoctoral Fellow, University of Oxford
  • Deirdre Mulligan — Panelist
    Associate Professor, School of Information, University of California at Berkeley
  • Michael Philips — Panelist
    Assistant General Counsel, Microsoft Corporation
  • Tom Yeh — Panelist
    Assistant Professor, Computer Science, University of Colorado Boulder

Know What’s Next