Recent technological change calls for a revolution in regulation.
Regulators around the world also are being challenged by questions about whether traditional legal concepts can apply to artificial intelligence (AI). What laws should apply? Who should enforce them? Who should be liable for what a self-teaching AI bot does? Solving these questions can remake administrative law, criminal law, and tort law. They show the perplexing and exciting challenges that young lawyers will confront throughout their careers.
I began this lecture by discussing the first administrative agency: the Interstate Commerce Commission (ICC). The ICC was born in a period of rapid technological advancement and innovation. More recent technology made the ICC obsolete, and it was abolished by Congress in 1995.
In the 1950s, the political scientist Marver Bernstein compared the “the rhythm of regulation” to the human life cycle, with phases of “gestation, youth, maturity, and old age.” Today, this life cycle is moving at a much more rapid pace than ever before. Administrative agencies need to respond to today’s innovations and advancements, including to the ways that technology affects agencies’ own ability to operate effectively.
The law school experience today is also dramatically different than it was years before. Students today have cell phones and tablets that allow them to avoid carrying legal casebooks around as I had to do.
In the same way, administrative agencies must come to terms with and understand the new world in which they operate. Not doing so would be the equivalent of failing to understand how the railroad was reshaping society 150 years ago.
The advent of digital technologies and the algorithms that power them will challenge the focus, tools, and techniques of administrative agencies. AI will raise a host of novel questions. Can an algorithm be completely transparent? How can regulation be effective without accountable people or paper? Are there any tools that agencies can reasonably obtain to deal with the new world?
We know that algorithms can be secretive. They can malfunction. They seldom declare their intent. And they can, as some like to say, “hallucinate”—or lie.
How do these characteristics inform public officials’ decisions about the right regulatory approach? In its impact, the AI era is bringing on a new “railroad revolution” multiplied many times over. How do regulators gain the understanding of technology necessary to craft regulation for 2023 and beyond?
Our regulatory paradigm is not ready for ChatGPT—or of the algorithms behind its actions—but these algorithms will revolutionize our agencies and how we regulate.
This is a particularly fascinating and exciting time to be studying and thinking about our regulatory paradigm in the United States. We have been down similar roads before. But there are new challenges that lie ahead.
Soon it will be the turn of the next generation of lawyers. It will not be long before those who are in law school today will be in positions of leadership and policymaking authority. And every single one of you attending or reading this lecture will be part of the upcoming revolution in regulation—whether you are affiliated with an administrative agency, a technology firm, an innovator, or as an attorney at a law firm.
What is your approach? Where do you draw the lines? It is you who will create Regulation 2.0.
This essay is part of a three-part series based on remarks delivered at the Annual Distinguished Lecture on Regulation at the University of Pennsylvania Law School on March 29, 2023.
A fully formatted version of this entire three-part series is also available for download as a single, integrated PDF article.