About | HeinOnline Law Journal Library | HeinOnline Law Journal Library | HeinOnline

LOG IN
LOG IN


92 Geo. Wash. L. Rev. 1473 (2024)
AI Regulation Has Its Own Alignment Problem: The Technical and Institutional Feasibility of Disclosure, Registration, Licensing, and Auditing

handle is hein.journals/gwlr92 and id is 1587 raw text is: At Regulation Has Its Own Alignment Problem: The Technical and Institutional Feasibility of Disclosure, Registration, Licensing, and Auditing Neel Guha, Christie M. Lawrence, Lindsey A. Gailmard, Kit T Rodolfa, Faiz Surani, Rishi Bommasani, Inioluwa Deborah Raji, Mariano-Florentino Cutllar, Colleen Honigsberg, Percy Liang & Daniel E. Ho* ABSTRACT Calls for regulating artificial intelligence (Al) are widespread, but there remains little consensus on both the specific harms that regulation can and should address and the appropriate regulatory actions to take. Computer scientists propose technical solutions that may be infeasible or illegal; lawyers propose regulation that may be technically impossible; and commentators propose policies that may backfire. AI regulation, in that sense, has its own alignment problem, in which proposed interventions are often misaligned with societal values. This Article assesses the alignment and technical and institutional feasibility of four dominant proposals for AI regulation in the United States: disclosure, registration, licensing, and auditing. The caution against the rush to heavily regulateAI without addressing regulatory alignment is underpinned by threearguments.First,Alregulatoryproposalstendtosufferfrom both regulatory mismatch (vertical misalignment) and value conflict (horizontal misalignment). Clarity about a proposal's objectives, feasibility, and impact may reveal that it is poorly matched with the harm intended to be addressed. In some instances, the impulse forAl regulation may, in fact, be better addressed by non-Al regulatory * Guha (lead author) is a ID. Candidate at Stanford Law School and Ph.D. Candidate in Computer Science at Stanford University. Lawrence (lead author) earned her ID. at Stanford Law School. Gailmaird is a Postdoctoral Scholar at the Regulation, Evaluation, and Governance Lab at Stanford Law School. Rodolfa is a Research Director at the Regulation, Evaluation, and Gover- nance Lab at Stanford Law School. Surani is a Research Fellow at the Regulation, Evaluation, and Governance Lab at Stanford Law School. Bommasani is a Ph.D. Candidate in Computer Science at Stanford University. Raji is a Ph.D. Candidate in Computer Science at University of California, Berkeley. Cu6llar is the President of the Carnegie Endowment for International Peace and a Visit- ing Scholar at Stanford Law School. Honigsberg is the Associate Dean of Curriculum and Profes- sor of Law at Stanford Law School. Liang is an Associate Professor of Computer Science at Stan- ford University. Ho (corresponding author, dho@law.stanford.edu) is the William Benjamin Scott and Luna M. Scott Professor of Law, Professor of Political Science, Professor of Computer Science (by courtesy), Senior Fellow at the Stanford Institute for Human-Centered Artificial Intelligence (HAI), Senior Fellow at the Stanford Institute for Economic Policy Research, and the Director of the Regulation, Evaluation, and Governance Lab, Stanford University. We thank HAI for support, Rui-Jie Yew for research assistance, and Michael Abramowicz, David Freeman Engstrom, Alicia Solow-Niederman, Russell Wald, and attendees of The George Washington Law Review's Annual Symposium in 2023 for helpful comments. December 2024 Vol. 92 No. 6 1473