Pending Congressional Revision of Section 230, Courts Should Treat it Like the Sherman Act
by Michael C. Dorf
In 2021, a panel of the U.S. Court of Appeals for the Ninth Circuit construed Section 230 of the Communications Decency Act mostly to shield major internet companies from civil liability to the families of victims of ISIS murders. The plaintiffs alleged that the algorithms of Facebook, YouTube, and Twitter promoted ISIS content, thus rendering them liable under the civil liability provision of the Anti-Terrorism Act (ATA). On Tuesday and yesterday of this week, the Supreme Court heard oral arguments in the two cases that seek reversal of the Ninth Circuit decision.
Why two cases? Because the Ninth Circuit decision reached a split decision. It approved one district court's determination that Section 230 provides a shield but reversed another district court's dismissal of the complaint on the ground that it failed to state an ATA claim. Accordingly, in Tuesday's argument in Gonzalez v. Google, the plaintiffs appealed, arguing that Section 230 should not be construed to shield platforms whose algorithms "decide" to display ISIS (and similar) content. In yesterday's argument in Twitter v. Taamneh, the Court heard Twitter's appeal of the portion of the Ninth Circuit's decision that found that the plaintiff had stated a claim under the ATA.
As a shorthand, think of the Twitter case as involving the construction of the ATA, while Gonzalez involves Section 230. To my mind, Gonzalez is the more important case because it is possible for the plaintiff to prevail in SCOTUS in Twitter but still lose the case on remand on the basis of the Section 230 shield. Gonzalez is also the more important case in that Section 230 provides a cross-cutting defense against all sorts of liability, not just liability under the ATA.
In the balance of today's essay, I'll offer an observation about the Gonzales oral argument to tee up a pitch for the proposition that titles the essay.
Relatively early in the Gonzales argument, Justice Kagan identified the heart of the problem: Section 230 was written mostly with internet service providers in mind at a time before social media (and other) platforms started using algorithms to display content to users based on what the users had previously chosen to view. But given the ubiquity of such algorithms today, a rule that renders the protection of Section 230 inapplicable to algorithm-driven content provision would render the protection virtually useless. The question with which the Justices and the lawyers (especially the plaintiff's lawyer) struggled was how to apply Section 230 in an algorithm-driven world.
Given the supposed textualist commitments of many of the Justices, I was surprised that only a small portion of the effort to answer that question focused on the text of Section 230. It provides: "No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider." On its face, that's pretty categorical and seems to favor the tech companies as against the ATA plaintiffs. Facebook, Google (through YouTube), and Twitter are "interactive computer service[s]" on whose platforms ISIS and other bad actors place despicable and harmful material, but, per Section 230, the platforms are immune from whatever liability would attach to a "publisher or speaker."
Social media sites surely count as interactive computer services, even though social media didn't exist when Congress enacted Section 230. However, if the algorithms of a social media site prioritize certain sorts of content in virtue of a user's prior browsing history, its virality, and/or other factors, does that mean (as the plaintiffs argue) that the content is no longer being "provided by another information content provider" or at least no longer entirely by another provider? In other words, does the operation of the platforms' algorithms make the platforms themselves content providers and thus unshielded by Section 230?
The text of Section 230(f)(4) seems relevant. It defines an eligible platform as including software that does
any one or more of the following:
(A) filter, screen, allow, or disallow content;
(B) pick, choose, analyze, or digest content; or
(C) transmit, receive, display, forward, cache, search, subset, organize, reorganize, or translate content.
It would be odd to say that YouTube qualifies for Section 230 protection in light of the fact that its algorithms "pick, choose, . . . display . . . organize [and] reorganize . . . content" but that it loses that protection by itself becoming the content provider rather than the provider solely of third-party content in light of the fact that its algorithms pick, choose, display, organize, and reorganize third-party content.
To be sure, the text of Section 230 is not necessarily dispositive. As Justice Kagan's initial observation implied, Congress wasn't thinking about algorithm-driven social media sites when it enacted Section 230, and there are enough textual ambiguities to support a range of results. Thus, while the Court could write an opinion that purports to parse the statutory text closely and to derive a result from that text, that would hardly be the best way--or even necessarily an honest way--to resolve the case.
And that brings me to my proposal. Ideally, Congress would revise Section 230 to take account of the technological changes and advances since it was enacted. Even if it did, however, given the pace of change in communications technology, new uncertainties and ambiguities would quickly arise. Congress could mitigate that risk by delegating to the FCC or some other federal agency or agencies substantial rule-making authority. Agencies can't act all that quickly, given the constraints of the Administrative Procedure Act, but they will typically act more quickly and more frequently than Congress.
So that's the first best option: new legislation modernizing Section 230 and delegating substantial regulatory authority to the FCC or other appropriate agencies. But there's no guarantee Congress will do that or do it soon. In the meantime, courts need to construe Section 230 as it exists. My suggestion is that they do so in roughly the same way that courts have construed the Sherman Act of 1890, which was supplemented by the Clayton Act in 1914 (and a few subsequent amendments and additions) but at its core has not been changed in over 130 years, despite enormous changes in the economy.
Consequently, courts have rightly treated the broad language of the antitrust laws as a grant of authority to fashion common law based on what some courts and scholars sometimes call an "economic constitution" for the country. I don't favor the "constitution" language, insofar as it suggests that Congress cannot amend the antitrust laws with a simple statute, but I agree with the sentiment that in light of the fact that Congress does not frequently amend the broad language of the antitrust laws, those laws cannot and should not be construed by closely parsing their text--which is highly under-determinate over a wide range of cases.
That is not to say that I agree with all of the particular ways in which courts have exercised their common-law-making authority with respect to antitrust policy. In particular, I'm highly dubious of the move in the 1970s and 1980s and under the influence of Robert Bork's work to reorient antitrust jurisprudence almost exclusively towards maintaining low prices for consumers. To my mind, the political economy concerns of progressives like Brandeis ought to have remained an important element of antitrust case law--not because the statutory text required that approach but because (as politicians across the political spectrum are rediscovering with respect to big tech) bigness and monopolization themselves carry risks apart from their impact on prices.
However, even as I disagree with the particular direction in which the courts have taken antitrust policy, I applaud their decision to treat the antitrust statutes as authorizing resort to policy considerations in common law fashion when confronting antitrust issues. The alternative of staring hard at the text of the Sherman, Clayton, and Robinson-Patman Acts to attempt to unearth meanings Congress did not place in those under-determinate texts is not viable.
Section 230 is substantially younger than the antitrust laws, but given the pace of communications technology innovation, it is nearly as under-determinate with respect to the algorithms of social media platforms as the landmark antitrust laws are with respect to contemporary competition policy. Faced with that reality, the courts are inevitably thrust into a policy making role with respect to the many questions left unanswered by the text of Section 230 and other statutes.
But wait. Are courts well positioned to make Internet policy? Here's the pushback from Justice Kagan during the Gonzalez argument: "we're a court. We really don't know about these things. You know, these are not like the nine greatest experts on the Internet."
The problem with that objection is that it's hardly clear which way it cuts. The Court has to decide the case one way or the other. Whatever it decides could make a mess of internet regulation. It has to make some judgment about what best promotes the policies Congress was pursuing in Section 230 and other statutes. Justice Kagan is right that ideally Congress would delegate that task to an expert agency, but unless and until it does so, the Court doesn't really have a choice but to be guided by its own admittedly limited best judgment.