On the (In)Applicability of Corporate Rights Cases to Digital Minds

Cross-posted from my per­sonal blog and AI Im­pacts.

High-Level Takeaway

The ex­ten­sion of rights to cor­po­ra­tions likely does not provide use­ful anal­ogy to po­ten­tial ex­ten­sion of rights to digi­tal minds.

Introduction

Ex­am­in­ing how law can pro­tect the welfare of pos­si­ble fu­ture digi­tal minds is part of my re­search agenda. I ex­pect that study of his­tor­i­cal efforts to se­cure le­gal pro­tec­tions (“rights”) for pre­vi­ously un­pro­tected classes (e.g., formerly en­slaved per­sons, non­hu­man an­i­mals, young chil­dren) will be cru­cial to this line of re­search.

I re­cently read We the Cor­po­ra­tions: How Amer­i­can Busi­nesses Won Their Civil Rights by UCLA con­sti­tu­tional law pro­fes­sor Adam Win­kler. The book chron­i­cles how busi­ness cor­po­ra­tions grad­u­ally won var­i­ous con­sti­tu­tional and statu­tory civil rights, cul­mi­nat­ing in the (in)fa­mous re­cent Ci­ti­zens United and Hobby Lobby cases.

A key in­sight from Win­kler’s book is that, con­trary to some pop­u­lar por­tray­als of cor­po­rate rights cases, these cases usu­ally do not rely pri­mar­ily on cor­po­rate per­son­hood: “While the Supreme Court has on oc­ca­sion said that cor­po­ra­tions are peo­ple, the jus­tices have more of­ten re­lied upon a very differ­ent con­cep­tion of the cor­po­ra­tion, one that views it as an as­so­ci­a­tion ca­pa­ble of as­sert­ing the rights of its mem­bers.” Id. at xx. The Court, in other words, “pierced the cor­po­rate veil” to give cor­po­ra­tions rights prop­erly be­long­ing to its mem­bers. See id. at 54–55.

The Supreme Court’s opinion in Ci­ti­zens United is illus­tra­tive. In de­ter­min­ing that the First Amend­ment’s free speech pro­tec­tions ap­plied to cor­po­ra­tions, the Court wrote: “[Un­der the challenged cam­paign fi­nance statute,] cer­tain dis­fa­vored as­so­ci­a­tions of cit­i­zens—those that have taken on the cor­po­rate form—are pe­nal­ized for en­gag­ing in [oth­er­wise-pro­tected] poli­ti­cal speech.” 558 U.S. at 356. The Court held that this was im­per­mis­si­ble: the share­hold­ers’ right to free speech im­bued the cor­po­ra­tion—which it viewed as merely an as­so­ci­a­tion of rights-bear­ing share­hold­ers—with those same rights. See id. at 365.

Ear­lier cases that, Win­kler ar­gues, ex­hibit this same pat­tern in­clude:

  1. Bank of U.S. v. De­veaux, hold­ing that fed­eral ju­ris­dic­tion over cor­po­ra­tions de­pends on ju­ris­dic­tion over the in­di­vi­d­u­als com­pris­ing the cor­po­ra­tion;

  2. Trus­tees of Dart­mouth Col­lege v. Wood­ward, hold­ing that cor­po­rate char­ters gave trustees pri­vate rights therein, which were pro­tected against state al­ter­a­tion by the Con­sti­tu­tion;

  3. NAACP v. Alabama ex rel. Pat­ter­son, hold­ing that non-profit cor­po­ra­tion could as­sert First Amend­ment rights of its mem­bers;

  4. Bains LLC v. Arco Prod­ucts Co., hold­ing that a cor­po­ra­tion had stand­ing to bring racial dis­crim­i­na­tion claim for racial dis­crim­i­na­tion against its em­ploy­ees.

Implications

I be­lieve that this un­der­stand­ing of the cor­po­rate civil rights “strug­gle” has small-but-non­triv­ial im­pli­ca­tions for a po­ten­tial fu­ture strat­egy to se­cure le­gal pro­tec­tions for digi­tal minds. Speci­fi­cally, I think Win­kler’s the­sis sug­gests that the ex­ten­sion of rights to cor­po­ra­tions is not a use­ful his­tor­i­cal or le­gal anal­ogy for the po­ten­tial ex­ten­sion of rights to digi­tal minds. This is be­cause Win­kler’s book demon­strates that cor­po­ra­tions gained rights pri­mar­ily be­cause their con­sti­tu­tive mem­bers (i.e., share­hold­ers) already had rights. In the case of digi­tal minds gen­er­ally, I see no ob­vi­ous anal­ogy to share­hold­ers: digi­tal minds as such are not mere ex­ten­sions or as­so­ci­a­tions of en­tities already bear­ing rights.

More con­cretely, this sug­gests that se­cur­ing le­gal per­son­hood for digi­tal minds for in­stru­men­tal rea­sons is not likely, on its own, to in­crease the like­li­hood of le­gal pro­tec­tions for them.

Car­rick Flynn sug­gested to me (and I now agree) that non­hu­man an­i­mal pro­tec­tions prob­a­bly provide the best ana­log for fu­ture digi­tal mind pro­tec­tions. To the ex­tent that it rules out an­other pos­si­ble method of ap­proach­ing the ques­tion, this post sup­ports that the­sis.

This work was fi­nan­cially sup­ported by the Berkeley Ex­is­ten­tial Risk Ini­ti­a­tive.