Artificial Intelligence and the Administrative State

As financial companies have begun employing automated advisors aimed at helping customers manage their money, and oncologists have started using the artificial intelligence system known as Watson to identify new treatments, the prominent role that sophisticated computer programs have begun to occupy in our lives has become undeniable. Government agencies are also harnessing the powers of automation. The U.S. Environmental Protection Agency and the U.S. Food and Drug Administration, for example, are starting to use complex computer models that can predict environmental exposure to chemicals and drug interactions across patient groups. As agencies begin to enter this brave new world of automation, questions have begun to emerge about how government officials should delegate important tasks to machines. For one, will automation negatively affect the level and quality of human deliberation and dialogue that are integral to democratic governance? Furthermore, when adjudicating individual determinations, like the awarding of disability benefits, will machines prove incapable of providing much needed empathy to claimants? And given recent investigations over cyberattacks occurring during the 2016 presidential election, what cybersecurity risks, if any, might an automated administrative state face? Mariano-Florentino Cuéllar, an Associate Justice on the California Supreme Court, addressed many of these issues in his keynote address delivered earlier this year at the University of Pennsylvania Law School, as part of the Penn Program on Regulation’s annual regulation dinner. This four-part series, Artificial Intelligence and the Administrative State, draws on Justice Cuéllar’s remarks delivered at the regulation dinner, where he presented his insights about the prospects—and potential pitfalls—of government use of machine-learning algorithms. Justice Cuéllar’s first essay addresses new digital technologies that can play a role in the regulatory process, as well as the dilemmas that these technologies may create. His second essay concerns the increasing use of automation in the administrative state, while his third essay outlines risks stemming from cyberdelegation. In his concluding essay, Justice Cuéllar offers solutions for addressing the challenges associated with administrative automation. The Regulatory Review is honored to have had Justice Cuéllar visit Penn Law, and, now, to have the privilege of featuring his incisive commentary and analysis on this critical topic. And we hope, in turn, that Justice Cuéllar’s insights will guide future automation efforts within the administrative state. December 19, 2016 | Mariano-Florentino Cuéllar Agencies can rely on sophisticated computer programs—programs that agencies could use to make or support decisions, and could therefore assume an increasingly prominent role in the regulatory process. In the coming years, computing power and storage will almost certainly become even cheaper than it is today, surveillance more pervasive, software architecture more flexible, and the limitations of human decision makers will become more salient. December 20, 2016 | Mariano-Florentino Cuéllar Even thoughtful experts who are familiar with the subtleties of environmental, national security, or public health data may fail to recognize patterns that can give agencies useful ideas about how to achieve their responsibilities. It is certainly understandable, then, why societies could become interested in making greater use of computer systems that hold the promise of improving the quality and integrity of administrative decisions. December 21, 2016 | Mariano-Florentino Cuéllar Thoughtful use of computers in administrative government—and in particular deployment of artificial intelligence technologies involving expert systems and deep learning—have the potential to increase consistency in decision-making and to help agency officials understand a complex world to make better decisions. But the advantages of cyberdelegation in the administrative state will also bring with them at least four sets of challenges warranting careful scrutiny. December 22, 2016 | Mariano-Florentino Cuéllar From cybersecurity risks to changes in public deliberation, government agencies’ use of automation and artificial intelligence will pose numerous challenges for the administrative state. Although no simple rubric exists to decide how to navigate these uncharted waters, there are a few possibilities for how agencies, policymakers, and the courts could increase society’s capacity to make informed choices about automation in the administrative state