Skip to content

Symposium: A Skeptical View of Information Fiduciaries


Lina Khan is an Academic Fellow at Columbia Law School.

David Pozen is the Charles Keller Beekman Professor of Law at Columbia Law School and the author of The Constitution of the War on Drugs.

This post introduces a symposium on Lina Khan and David Pozen’s essay “A Skeptical View of Information Fiduciaries.” Read the rest of the posts here.

*     *     *


In recent years, the concept of “information fiduciaries” has surged to the forefront of debates on platform regulation. Developed by Professor Jack Balkin, the informationfiduciary proposal seeks to mitigate the asymmetry of power between a handful of dominant digital firms and the millions of people who depend on them. Just as doctors, lawyers, and accountants are assigned special legal duties of care, confidentiality, and loyalty toward their patients and clients, Balkin argues that Facebook, Google, and Twitter should owe analogous duties toward their end users. This argument has gained broad support. Last December, over a dozen Democratic Senators introduced legislation that would designate online service providers as fiduciaries for their users, effectively implementing Balkin’s proposal.

In a forthcoming essay, we question the wisdom of applying a fiduciary framework to dominant digital platforms. Focusing on the case of Facebook—Balkin’s central example of a purported information fiduciary—we identify a number of lurking tensions in the proposal. For instance:

  • Unlike most doctors and lawyers, the directors of publicly traded firms such as Facebook already owe fiduciary duties to the firms’ shareholders. Creating a new set of user-facing fiduciary duties could lead to deeply divided loyalties, because what is good for Facebook’s users will often be bad for shareholders. Moreover, it is hard to see how this tension could be reconciled in favor of users without dismantling the company’s behavioral-advertising-based business model—a step Balkin has explicitly ruled out. So long as Facebook’s only money-paying customers are advertisers, it will have little economic incentive to place users’ interests first.
  • Unlike doctors and lawyers, who perform specialized tasks and provide individualized judgments, Facebook does not bring comparable professional expertise to bear on behalf of customers. Rather, it supplies them with access to a communications network. While doctors and lawyers must collect a certain amount of sensitive information to be able to serve their patients and clients well, sharing reams of personal data with Facebook (let alone permitting Facebook to track you across the internet) is not a functional prerequisite to accessing a communications network. It is the price Facebook has chosen to set.
  • In traditional fiduciary settings, the beneficiary understands and approves the core terms of the relationship. Yet surveys suggest that over two-thirds of Facebook’s users don’t even know that the company collects data to classify their interests and traits. An overwhelming majority of users, moreover, do not want to receive any personalized advertisements—which, again, are the engine of Facebook’s business. The principle of informed consent arguably breaks down under these conditions.

One of the major selling points of the information-fiduciary approach, on Balkin’s account, is that it would help laws and policies adopted in its name to survive “neo-Lochnerian” First Amendment review. A closer analysis of the Roberts Court’s free speech jurisprudence, however, suggests that this constitutional payoff is probably chimerical. Nor is it clear why anti-Lochnerians would find fiduciary logic an attractive basis for justifying government regulation in this area.

Against this seemingly illusory First Amendment benefit, the potential costs of adopting an information-fiduciary framework are nontrivial. In particular, the framework risks painting a false portrait of the digital world and thereby inviting policy misfires. It is Facebook’s business model coupled with its market dominance—not the company’s unwillingness to act as a loyal data caretaker—that is the key source of many of the harms that demand a legal response. By failing to grapple with the structural dimensions of platform power, the information-fiduciary proposal is liable to cannibalize rather than complement necessary reforms.

Or so we argue. The draft of our essay posted on SSRN is a true draft, subject to significant revision. We are grateful to the LPE Blog editors for hosting this symposium on our paper, and we look forward to learning from all of the contributors.