Governing Algorithms and Algorithmic Governance
Course Numbers: INST878D/CMSC839C
The creation of modern AI/ML systems has created fundamentally new regulatory dilemmas and exacerbated long-standing regulatory challenges for computing technologies. These stem from the fact that technologies are never simply “good” or “bad,” but always possess the potential for both beneficial and harmful application—and AI/ML systems appear to promise extreme benefits and pose extreme risks. Cognisant of this challenge, technical researchers, policy experts, technology companies, governments and other stakeholders around the world are racing to answer the following question:
What social governance and technical mechanisms can we use to maximize the positive societal impact of AI/ML systems and minimize the risks associated with these systems?
Ongoing efforts to answer this question have required close collaboration between many experts, including technical practitioners with deep understanding of the ways in which AI/ML systems operate, policy experts that understand the ways in which specific policy choices are likely to shape technological deployments, and domain-specific experts that can document the benefits and harms induced by these technologies. The result of this collaborative effort is a slate of detailed technical research, fascinating multi-stakeholder reports, risk management frameworks, critical academic studies, newly proposed regulations or policy approaches—including in the federal, state, and international regimes—and statements of values. Despite this progress, the rapid rate at which AI/ML systems are changing and growing means that there are always new challenges to address.
This course brings the study of this critical question into the classroom in a multi-disciplinary way that mirrors that approach currently being used in practice. The course (see draft schedule below) is divided into three units: (1) Setting the stage, in which we will explore the immense stakes of making good regulatory and policy decisions for AI/ML systems and look at the emerging policy proposals; (2) Core existing challenges, in which we explore some of the better understood challenges and the technical state-of-the-art when it comes to addressing them; and (3) Emerging challenges, in which we look at the bleeding edge challenges from both a technical and policy perspective. Given the rapid rate at which this area continues to evolve, we expect that the exact topics and readings covered will likely change in response to breaking news during the semester.
This course counts as "qualifying" for CS PhD students within the AI area.
Instructors
Lee J Tiedrich
AIM Fellow
College of Information Studies
Email: tiedrich (at) umd (dot) edu
Gabriel Kaptchuk
Assistant Professor
Department of Computer Science
Email: kaptchuk (at) umd (dot) edu
Rachel Thomas (Teaching Assistant)
Department of Computer Science
Email: rthomase (at) umd (dot) edu
Course Information
Syllabus and Grading: Available upon request.
Meeting Times: Tuesdays, 11:00am–1:45pm
Course Structure: Guest lectures, disccussion-based, with many in-class activities.
Course website: ga-ag.com
Course Project: Available Here