From Law to Automation

Thomas A. Smith

Purchase a reprint version of the Article (Amazon) | Read the Article (PDF) | Download the Article (PDF) Download the Article (PDF)

Laws are like computer code in that they describe what must be done or not done. Computer code does the same thing, but in a sense more precisely. A law can provide that “no vehicle shall be allowed in the park.” A computer controlling the town’s robot park rangers can do the same thing, by stopping all vehicles from entering the park. The law might define “vehicles” to exclude ambulances, perambulators, and police cars. The robot rangers can also be programmed to let these non-vehicles pass.

But in a sense, law is not like code. This is because law has human agency as an essential step in the causal chain between its enactment and its enforcement or performance. In a seeming paradox, one must be able to disobey a rule before that rule is a law. With computer code or any automatic process, this is not the case. In fact, a process is automated to the extent human agency is removed from it.

Computers do what they are programmed to do. They don’t make mistakes; they are “just machines,” at least so far. Computer programmers often make mistakes. The business of writing code is fearsomely complicated, all but assuring there will be bugs to be eliminated from a program, and a program is never final. In the end, though, this mechanical quality makes computers, like any machine, incapable of being moral. They are non-moral. Machines or automatons cannot be wrong in the normative sense. The non-culpable “mistakes” that machines make are made because they were wrongly programmed. If blame is to be assigned, it must be assigned to the humans ultimately responsible for getting a particular task done.

My thesis in this article is that more and more of what is now done by law in moral space by relying on human agency is being pushed into non-moral space as these objects of regulation are subjected to mechanisms that are not social, but purely mechanical. In other words, automation occurs. Automation involves the disappearance of law. If a process becomes purely mechanical, it is no longer a legal process. A legal process has somewhere in it at least one step that involves human agency.

Volume 1
Read the Article (PDF) Flip through the Article

Cite as

Thomas A. Smith, From Law to Automation, 1 Criterion J. on Innovation 535 (2016).

Thomas A. Smith is Professor of Law at the University of San Diego School of Law. He received an A.B. in philosophy and history from Cornell University, where he was a Telluride Scholar, and then received a B.A. (First Class) in politics, philosophy, and economics from Oxford University, where he was a Rhodes Scholar. He received a J.D. from Yale Law School, where he was Notes Editor of the Yale Law Journal. Professor Smith was a law clerk to Judge George MacKinnon on the U.S. Court of Appeals for the D.C. Circuit and later served as Senior Counsel on President Reagan’s Council of Economic Advisers. He practiced international corporate law with Covington & Burling before joining the University of San Diego faculty in 1992. His scholarship addresses constitutional law, corporate and business law, and intellectual property law. Professor Smith has published his research in the Yale Law Journal, the California Law Review, the Michigan Law Review, the Northwestern University Law Review, and other journals.