The Interesting Robot: A Reply to Professor Abbott
Ryan Calo · Jerusalem Review of Legal Studies · 2022
I had a philosophy professor in college, Bob Fogelin, who sorted classical arguments into four categories: (1) interesting and right, (2) interesting and wrong, (3) uninteresting and right, and (4) uninteresting and wrong. Very few arguments that stand the test of time fall in the last category. Even fewer fall into the first.1 Most worthwhile work falls somewhere in the middle. Ryan Abbott has written a wonderful, interesting book about artificial intelligence and the law that happens to be wrong. The book is interesting (and ambitious) in that Abbott manages to articulate a nuanced conceptual framework for AI that spans at least four distinct legal contexts. Along the way, he makes any number of fascinating and trenchant observations regarding the interaction of law and technology. The book is wrong, or at least incomplete, insofar as it argues for a concept—“AI legal neutrality”—without satisfactory criteria of application. The Reasonable Robot begins by describing a puzzle of utmost importance to AI and the law: how should policymakers and courts think about harms or benefits generated by the substitution of machines for the activities of people? Rather than focus on a single context such as transportation or war, Abbott traces this question across numerous sectors and legal fields, from taxation to tort law to intellectual property to crime. A prolific scholar, Abbott has written about the impacts of AI in each of these contexts. But the book does far more than summarize prior work. Abbott ties these disparate areas together with a unified thesis.