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.