On 21 April this year, the European Commission published its proposal for a Regulation laying down harmonised rules on AI—the first comprehensive draft legislative package dealing specifically with AI systems across multiple sectors. The NRCCL is organizing a set of seminars in which we subject the Commission’s proposal—and policy discourse around it—to closer scrutiny. These seminars take place under the aegis of the ongoing VIROS project (‘Vulnerability in the Robot Society’) funded by the Research Council of Norway and run by the NRCCL/SERI in cooperation with a network of other institutions and researchers. The seminars are gratis and conducted on zoom.
In the initial seminar, held 25 May, we considered the origins, rationale, and basic mechanics of the Commission’s proposal. In upcoming seminars, we focus on the proposal in a global context.
The first of these upcoming seminars considers the proposal from US perspectives. Amongst the issues taken up are the degree of compatibility between the Commission’s proposal and US regulatory policy, the pros and cons of core elements of the proposal in light of US thinking, and whether we can expect some sort of ‘Brussels Effect’ in AI regulation, vis-à-vis the USA.
Presentations
The seminar features presentations by Margot Kaminski and Dennis D. Hirsch, each of whom perceives AI regulatory policy in the US as manifesting a turn away from individual rights and towards other regulatory approaches. Hirsch will discuss the emerging embrace of substantive requirements (e.g. duties of care, or unfairness determinations). Kaminski will address the increased reliance on systemic, procedural requirements (e.g. impact assessments). Each will compare US policy trends with corresponding regulatory proposals in Europe.
Dennis D. Hirsch is Professor of Law at both the Ohio State University Michael E. Moritz College of Law and at Capital University Law School. At Ohio State, he directs the Program on Data and Governance, a program conducting non-partisan research on the governance of advanced analytics and AI. His current work focuses on the new legal paradigms and organizational practices required to promote fair, just, and transparent AI. In 2010, he served as a Fulbright Senior Professor at the University of Amsterdam’s Institute for Information Law (IViR) and continues to teach in the IViR Summer Course on Privacy Law and Policy.
Margot Kaminski is Associate Professor of Law at the University of Colorado and the Director of the Privacy Initiative at Silicon Flatirons. She specializes in the law of new technologies, including AI, robots, and drones (UAS). In 2018, she researched comparative data privacy law in the Netherlands and Italy as a recipient of the Fulbright-Schuman Innovation Grant.
Programme
16:00–16:05 Welcoming address by Lee A. Bygrave (NRCCL)
16.05–16:25 Beyond notice-and-consent-based approaches to AI regulation:
common ground between US and European thinking?
Dennis D. Hirsch (Ohio State University and Capital University)
Advanced analytics and AI pose threats to people and society that traditional, notice-and-consent-based privacy law, cannot handle. Addressing these threats requires a new regulatory approach in which public authorities draw substantive lines between AI practices that are socially acceptable, and those that are not, regardless of whether individuals ‘consent’ to these uses of their data. Legal scholars and, increasingly, policymakers in the US have proposed regulatory regimes for the commercial use of AI that would draw such substantive lines. These proposals share some common ground with the European Commission’s proposed AI regulation. But are they completely compatible with the latter?
16:25–16:40 Discussion of Hirsch’s presentation
16:40–16:45 Break
16.45–17:10 Risk-based regulation versus individual rights: pros and cons for both sides of the Atlantic.
Margot Kaminski (University of Colorado)
In the United States, as in the European Commission’s proposed AI regulation, proposals to regulate AI have largely focused on systemic risk regulation, using regulatory tools such as impact assessments or audits. By contrast, the GDPR, the Council of Europe, the OECD and several other countries have emphasized the need for also including individual rights, such as the right to meaningful information about the logic involved and/or the right to contest an AI decision. What is gained by emphasizing the risk regulation approach—and what is lost by foregoing individual rights?
17:10–17:25 Discussion of Kaminski’s presentation
17:30 End of seminar
Welcome!