Content Moderation and the First Amendment

Back in 2013 I argued that under the Supreme Court’s jurisprudence algorithmic editing is speech for First Amendment purposes, and in Moody v. NetChoice the Supreme Court squarely so stated, in the context of social media platforms’ algorithmic prioritization of content.

Many people are unhappy about this jurisprudence, for understandable reasons. The invalidation of laws regulating websites’ substantive algorithmic decisions is a striking, and arguably unsettling, prospect. Some of the critics argue for revamping First Amendment jurisprudence. That would extend far beyond the treatment of editorial choices, but others respond more directly to this jurisprudence – for example, treating social media platforms as state actors or common carriers.

In my forthcoming Content Moderation and the First Amendment, I discuss these possible responses designed to limit Moody, concluding that the strongest candidate is excluding from First Amendment coverage editorial judgments made by monopolists. Unlike the other possible responses, it avoids arbitrary distinctions and focuses on companies that, by hypothesis, lack significant competitors.

I also discuss a prominent possibility that would expand the Moody line of cases – that audience interests alone are sufficient to trigger First Amendment coverage, so that messages sent by AI without any meaningful human involvement (and thus no traditional speaker for First Amendment purposes) are treated as speech.

As I discuss in the conclusion, my inclination is not to adopt either of these possibilities. As to market power, I err on the side of avoiding ad hoc exceptions. And the prospect of protecting speech generated entirely by AI seems sufficiently transformative that I favor caution.

I end by suggesting that these issues will become more fraught insofar as social media platforms become more influential and the prospect of artificial general intelligence becomes more real. And cleavages in response to those two developments will likely better explain individuals’ reactions to the treatment of algorithmic editorial decisions than more traditional First Amendment fault lines will. In this and in other areas, the ground is shifting beneath our feet, destabilizing current debates.