Algorithms

In her analysis of government speech, Seana Shiffrin reminds us that “government officials and licensed experts have heightened responsibilities, given their roles, of both sincerity and accuracy and often have special access to sources of information, thereby reducing opportunity for external verification and rebuttal” (p. 1025). This asymmetry in access to knowledge, paired with the authority these figures hold, gives their speech unique weight - and unique potential for harm when misused. Shiffrin rightly argues that such speech should be scrutinized under heightened standards because of the distinct way that it shapes public understanding.

I am curious as to what constitutional obligations algorithms (and their creators, i.e., the platforms that house them) should be held to. Specifically, the algorithms deployed by private social media platforms that govern what is seen, what is buried, and what circulates. These algorithmic systems, though ostensibly neutral tools, make decisions that can suppress, amplify, or reframe speech at scale. As such, they too obstruct the preconditions for meaningful public discourse.

To borrow Shiffrin’s own terms, algorithmic suppression and information funneling also “reduce opportunity for external verification and rebuttal” (p. 1025). When speech is pushed to the margins or made functionally invisible by opaque platform mechanisms, the possibility of critical engagement or correction (central to the First Amendment’s aims) is diminished. The structure of this suppression is different from the interpersonal interruptions we might experience in analog spaces. As she notes elsewhere, “unlike the role of the government, private companies do not serve the public as an end in itself” (p. 1016), and therefore are not bound by constitutional duties in the same way. But this legal distinction should not preclude a normative or political reckoning with the power these companies wield over our public sphere.

So for example: If Professor Hurley interrupts me during a classroom discussion, it is an infringement on my speech, but a relatively minor one. It doesn't meaningfully violate my First Amendment rights. However, if the algorithm running a platform where I speak ensures that my post is suppressed (effectively hidden from view by the very people it’s meant to reach) this becomes a much more serious incursion. It is not just about my inability to speak; it is about my inability to be heard. In this way, algorithmic suppression “obstructs the operation of the preconditions for achieving the First Amendment’s core purposes and in that way violates the First Amendment” (p. 1010), at least in spirit if not in strict legal doctrine.

One might argue, as Shiffrin hints, that because private platforms do not serve the public as an end in itself, their obligations should differ from those of the government. This is a valid counterclaim. After all, the First Amendment is traditionally interpreted as a shield against state action, not private decisions. But that interpretation assumes a world in which public discourse is evenly distributed across various private and public channels. That world no longer exists. Our media is dominated by a handful of private companies, companies that now effectively are the public square; their algorithmic choices have structural consequences for speech rights and democratic health. The algorithms in question are not passive conduits of information. They are active participants in shaping what we know, how we think, and which voices count. When designed to maximize engagement or profit, rather than truth or fairness, they tend to elevate inflammatory content and suppress nuance. They privilege repetition over rebuttal, and virality over verification (I was proud of my alliterations here btw). And they do so at scale, with little transparency and even less accountability.

If government officials are bound by heightened standards because of their authority and access to information, we should consider how platforms with outsized control over information flows might merit similar scrutiny. The logic that applies to public officials (namely, that their power demands responsibility) should extend, at least normatively, to private actors whose systems govern the very conditions under which speech is heard and understood.