Discretion in the Automated Administrative State

By Sancho McCann · , edited:

Sancho McCann, “Discretion in the Automated Administrative State” (2023) 36:1 Can JL & Jur 171.

Abstract

Automated de­ci­sion-mak­ing takes up an in­creas­ing­ly significant place in our ad­min­is­tra­tive state. This ar­ti­cle presents a con­cep­tion of dis­cre­tion that is help­ful for eval­u­at­ing the prop­er place of al­go­rithms in pub­lic de­ci­sion-mak­ing. I ar­gue that the al­go­rithm it­self is not a site of dis­cre­tion. The threat is that au­to­mat­ed de­ci­sion-mak­ing al­ters the re­la­tion­ships be­tween tra­di­tion­al ac­tors in a way that can cut down dis­cre­tion and hu­man com­mit­ment. Algorithmic de­ci­sion-mak­ers can serve to fet­ter the dis­cre­tion that the leg­is­la­ture and the pop­u­lace ex­pect to be ex­er­cised. We must strive to main­tain dis­cre­tion, moral agency, de­lib­er­a­tive ideals, and hu­man com­mit­ment through the sys­tem that sur­rounds the use of an al­go­rithm and to de­vel­op a new ex­per­tise that can re­tain and ex­er­cise the ex­pect­ed dis­cre­tion. Backing this ar­gu­ment are tra­di­tion­al le­gal con­straints, pub­lic ex­pec­ta­tions, and ad­min­is­tra­tive law prin­ci­ples, tied to­geth­er through the or­ga­niz­ing prin­ci­ple of dis­cre­tion.

Thanks

I would like to thank Professor Mary Liston for su­per­vis­ing this re­search project and for our many en­gag­ing dis­cus­sions. I also thank the par­tic­i­pants of the Algorithms and Rule of Law read­ing group con­vened at the Peter A. Allard School of Law.