31 Pages Posted: 22 Mar 2018 Last revised: 22 Feb 2019
Date Written: March 19, 2018
Automated legal systems occupy a central place in the administration of most regulatory regimes. Examples include TurboTax, wage and hour software, and self-driving cars. These systems produce results which invite examination and adjudication on a centralized, ex post basis. This is revolutionary. It means that the content of law, which technically applies to individual regulated parties, is determined centrally by interactions between the government and firms that make automated legal systems. I call this trend government-to-robot enforcement.
Government-to-robot enforcement could improve greatly the efficiency of compliance, which is vulnerable to underdetection and underenforcement. Government-to-robot enforcement could find mistakes, impose strict liability, use a damages multiplier, and resolve government claims against individual users collectively by dealing directly with the automated system. It could free innocent third parties from the negative externalities of noncompliance, like higher taxes because of others' tax avoidance or contaminated air because of illegal pollution. But government-to-robot enforcement has other problems. These include the risk of capture of government by an automated system, the risk of "reverse" capture of a system by government, the decline of individual claims against the government, and problems of winners and losers.
Keywords: enforcement, compliance, negative externalities, robot, cyberspace, automation, centralization, capture
JEL Classification: K23, K34, K42
Suggested Citation: Suggested Citation