Publications

The Committee on Standards in Public Life has recently published (February 2020) a review on ‘Artificial Intelligence and Public Standards’. Chaired by Lord Evans of Weardale KCB DL, the report takes a thorough look at the use of AI in public service through the framework of the Nolan Principles (Selflessness, Integrity, Objectivity, Accountability, Openness, Honestly, and Leadership). This paper briefly comments upon and analyses selections form the publication by surveying the recommendations.

An agency that fails to adequately consider the costs and benefits of its proposed regulatory changes increasingly places its rules at risk upon judicial review. Over the last couple of decades, courts have begun to expect agencies to use regulatory analysis techniques like cost-benefit analysis to justify their regulatory choices. This poses acute risks for the independent regulatory agencies, whose draft regulations are not reviewed by the Office of Information and Regulatory Affairs (OIRA), the White House experts on cost-benefit analysis. Independent regulatory agencies have not fully opted in to OIRA’s regulatory review, likely because they expect that OIRA review portends the end of their ability to make independent decisions. But what if it didn’t? This article draws upon the author's 10+ years at OIRA to offer a new way forward: bespoke regulatory review. That is, bilateral negotiations resulting in agreements between independent regulatory agencies and OIRA, custom fit to each agency’s unique features.

Scholars of regulation generally view the procedures that agencies must follow when promulgating rules as instruments by which political principals control bureaucratic agents. Much like political principals attempt to use procedural checks to constrain regulatory agencies actions, these same agencies employ various regulatory instruments to influence the decisions of private agents, especially firms. Despite the parallel nature of these principal-agent problems, few studies, if any, have looked at whether lessons from one can be used to inform the other. In this paper, we draw analogies between benefit-cost analysis (BCA)—a procedural control employed in the regulatory process—and three regulatory instruments that have similarities to BCA—performance standards, information disclosure requirements, and management-based regulation. We use lessons from research on the effectiveness of regulatory instruments to make predictions regarding the efficacy of BCA in various situations. Just as different regulatory instruments are appropriate for different regulatory contexts, the pathways by which BCA attempts to encourage better regulation may not all be applicable in every circumstance. We argue that such mutual exclusivity should inform how requirements for BCA are designed and that BCA’s emphasis on systematic analysis—the pathway most closely resembling management-based regulation—may offer the most promise for encouraging better rules.

Studi Parlamentari e di Politica Costituzionale, 203-204, 1-2

At every level of government, officials contract for technical systems that employ machine learning—systems that perform tasks without using explicit instructions, relying on patterns and inference instead. These systems frequently displace discretion previously exercised by policymakers or individual front-end government employees with an opaque logic that bears no resemblance to the reasoning processes of agency personnel. However, because agencies acquire these systems through government procurement processes, they and the public have little input into—or even knowledge about—their design or how well that design aligns with public goals and values. This Article explains the ways that the decisions about goals, values, risk, and certainty, along with the elimination of case-by-case discretion, inherent in machine-learning system design create policies—not just once when they are designed, but over time as they adapt and change. When the adoption of these systems is governed by procurement, the policies they embed receive little or no agency or outside expertise beyond that provided by the vendor. Design decisions are left to private third-party developers. There is no public participation, no reasoned deliberation, and no factual record, which abdicates Government responsibility for policymaking. This Article then argues for a move from a procurement mindset to policymaking mindset. When policy decisions are made through system design, processes suitable for substantive administrative determinations should be used: processes that foster deliberation reflecting both technocratic demands for reason and rationality informed by expertise, and democratic demands for public participation and political accountability. Specifically, the Article proposes administrative law as the framework to guide the adoption of machine learning governance, describing specific ways that the policy choices embedded in machine-learning system design fail the prohibition against arbitrary and capricious agency actions absent a reasoned decision-making process that both enlists the expertise necessary for reasoned deliberation about, and justification for, such choices, and makes visible the political choices being made. Finally, this Article sketches models for machine-learning adoption processes that satisfy the prohibition against arbitrary and capricious agency actions. It explores processes by which agencies might garner technical expertise and overcome problems of system opacity, satisfying administrative law’s technocratic demand for reasoned expert deliberation. It further proposes both institutional and engineering design solutions to the challenge of policymaking opacity, offering process paradigms to ensure the “political visibility” required for public input and political oversight. In doing so, it also proposes the importance of using “contestable design”—design that exposes value-laden features and parameters and provides for iterative human involvement in system evolution and deployment. Together, these institutional and design approaches further both administrative law’s technocratic and democratic mandates.