Many Fifth Circuit Judges Hope to Eviscerate Section 230–Doe v. Snap

I previously covered the district court ruling in this case. I summarized:

A high school teacher allegedly used Snapchat to groom a sophomore student for a sexual relationship. (Atypically, the teacher was female and the victim was male, but the genders are irrelevant to this incident).

The teacher was sentenced to ten years in jail, so the legal system has already held the wrongdoer accountable. Nevertheless, the plaintiff has pursued additional defendants, including the school district (that lawsuit failed) and Snap.

The lower court dismissed the claims against Snap per Section 230:

Doe’s negligent undertaking claim alleges that Snapchat has failed to monitor content and messages sent between parties on its platform. Doe’s gross negligence claim alleges that Snap is indifferent to sexual predators’ use of the platform to message minors “with apparent impunity.” Each of these claims seek to “fault[ ] [Snap] for information provided by [a] third part[y]”—messages and photos sent by Guess-Mazock to Doe. Section 230 provides Snap with immunity from these claims.

The plaintiffs invoked Lemmon v. Snap to get around Section 230. The lower court was unpersuaded: “Doe’s negligent design claim similarly aims to hold Snap liable for communications exchanged between Doe and Guess-Mazock.”

In June 2023, the Fifth Circuit affirmed the dismissal in a 3-page per curiam opinion. The court says simply that the plaintiff’s “argument is contrary to the law of our circuit” (cites to Doe v. MySpace and Diez v. Google). Nevertheless, the per curiam opinion ends ominously:

we are bound by the decisions of prior panels until such time as they are overruled either by an en banc panel of our court or by the Supreme Court

The plaintiff then sought en banc review. Yesterday, the Fifth Circuit denied the en banc request in a vote of 7 for review and 8 against review (with 2 judges not voting). Snap’s dismissal stands, but by a razor-thin margin.

The judges supporting review issued a dissent to the denial, written by Judge Elrod. As you might expect from the Fifth Circuit (where, as I’ve said often, the rule of law goes to die), the dissent opinion is painful reading filled with misinformation and shifting goalposts.

The dissent’s primary goal seems to be to urge the Supreme Court to take this case. In that vein, the dissent cites Justice Thomas’ anti-230 Malwarebytes statement SIX TIMES. As you may recall, Justice Thomas expressly acknowledged he wrote that statement without the benefit of briefing. Yet, to Justice Thomas stans, this non-precedential statement represents a more definitive analysis of Section 230 than the thousand-plus cases actually interpreting Section 230 over the past quarter-century, including the Fifth Circuit’s own long-standing precedent.

The dissent laments that the Fifth Circuit is not “revisiting our erroneous interpretation of Section 230, leaving in place sweeping immunity for social media companies that the text cannot possibly bear.” This means we’re going to get a textualist interpretation, accurate or not, but also we’re going to get policy and “living” text arguments because…why not?

The Textualist Argument

The dissent says:

Section 230 closes off one avenue of liability by preventing courts from treating platforms as the “publishers or speakers” of third-party content. Sub-section (c)(1) and (c)(2) say nothing about other avenues to liability such as distributor liability or liability for the platforms’ own conduct.

It’s mind-blowing that we’re still tussling about a publisher/distributor distinction in Section 230 in 2023. Zeran settled this issue in 1997, Congress expressly endorsed the Zeran interpretation in 2002 in the Dot Kids Implementation and Efficiency Act’s legislative history, and a quarter-century of jurisprudence has reaffirmed the irrelevance of the distinction. So are we doing it all over again? Why, yes we are:

Section 502 of the Communications Decency Act expressly authorizes distributor liability for knowingly displaying obscene material to minors. 47 U.S.C. § 223(d). This includes displaying content created by a third-party. It strains credulity to imagine that Congress would simultaneously impose distributor liability on platforms in one context, and in the same statute immunize them from that very liability.

[Reminder: Section 230 was section 509 of Telecommunications Act of 1996].

FFS. This kind of textualist “gotcha” is mockable. It’s a well-known story that Congress smashed together two competing bills to create the CDA without trying to harmonize them. For more, see Prof. Kosseff’s guest blog post on the legislative history (and his many other pieces on this subject). So the CDA’s textual inconsistency only “strains credulity” if one completely ignores the legislative manufacturing process that created the CDA.

The dissent then says the Fifth Circuit’s 15-year-old Doe v. MySpace precedent diverges from the Ninth Circuit’s Lemmon v. Snap decision, which allows some negligent design claims to bypass Section 230:

Immunity from design defect claims is neither textually supported nor logical because such claims fundamentally revolve around the platforms’ conduct, not third-party conduct. Nowhere in its text does Section 230 provide immunity for the platforms’ own conduct. Here, Doe brings a design defect claim. He alleges that Snap should have stronger age-verification requirements to help shield minors from potential predators. He further alleges that because “reporting child molesters is not profitable,” Snap “buries its head in the sand and remains silent.” Product liability claims do not treat platforms as speakers or publishers of content. “Instead, Doe seeks to hold Snap liable for designing its platform to encourage users to lie about their ages and engage in illegal behavior through the disappearing message feature.”

So, this is all misleading at best. The Ninth Circuit made clear that Lemmon v. Snap survived Section 230 preemption only because the plaintiffs were not suing over third-party content. But in this case, no matter how the plaintiff frames the complaint, the alleged legal violations necessarily turn on third-party content, i.e., the teacher’s messages to the student. Take away those messages, and the plaintiff’s negligent design claim would be nonsensical. For example, Snap’s allegedly inadequate age authentication (which would be unconstitutional for the court to impose, BTW) would be inconsequential if the student never accessed or received third-party content. The Snap app would be useless but harmless to this plaintiff without the third-party content. In contrast, in the Lemmon case, the plaintiffs argued that the app motivated the users to engage in dangerous offline behavior regardless of any third-party content. By eliding the critical role that third-party content plays to this plaintiff’s claims, the dissent creates and then attacks a legal fiction.

The stakes are high here. Every plaintiff can find a way to argue “negligent design” to work around Section 230, i.e., the service always could have chosen a different design to prevent the plaintiff’s harm, even if the harm was caused by third-party content. So the dissent’s misinterpretation of Lemmon v. Snap would completely nullify Section 230. That’s why the Lemmon v. Snap opinion was so precise about restricting the plaintiff’s claims to content that didn’t come from third parties, and why this dissent’s mischaracterization is so pernicious.

The Policy Analysis

Having beaten down its strawman argument using a textualist reconstruction, the opinion switches to a policy analysis, calling Section 230 “unmoored…from the background legal principles against which it was enacted.” So, these Fifth Circuit judges are sometimes interested in non-textualist considerations….? The dissent essentially reiterates its textualist analysis, because bogus arguments magically become convincing if they are repeated twice.

The “Living” Text Analysis

The dissent then turns to the “needs of a changing world.” This is the opposite judicial interpretation technique of textualism. Indeed, Constitutional originalists emphatically reject the “needs of a changing world” when interpreting text.

Continuing to bash Doe v. MySpace, the dissent says “our atextual transformation of Section 230 into a blunt instrument conferring near-total immunity has rendered it particularly ill-suited to the realities of the modern internet.” Smells like…judicial activism.

The dissent then offers this passage:

As the internet has exploded, internet service providers have moved from “passive facilitators to active operators.” They monitor and monetize content, while simultaneously promising to protect young and vulnerable users. For example, Snap itself holds itself out to advertisers as having the capability to target users based on “location demographics, interests, devices . . . and more!”

Today’s “interactive computer services” are no longer the big bulletin boards of the past. They function nothing like a phone line. Rather, they are complex operations offering highly curated content.

I wrote these bulleted critiques through heaving sobs:

  • Services were never passive facilitators. That’s 100% #STRAWMAN. Congress overturned Stratton Oakmont so that services, like Prodigy, would NOT act as passive facilitators.
  • Monitoring content and promising to protect young and vulnerable users is literally what Section 230 sought to advance–because Congress worried the services wouldn’t do this socially important work without the immunity that the dissenting Fifth Circuit judges seek to strip away.
  • Of course services don’t function like a phone line. They never have, no one ever would want them to do so, and the Supreme Court held in 1997 that services aren’t phone companies.
  • Section 230 enables services to offer highly curated content. As Wyden and Cox explained in their FCC comments, THAT WAS ALWAYS CONGRESS’ GOAL.

I have no idea what the dissent thought this passage accomplished, other than to repackage some legally inconsequential gripes.

Finally, nowadays partisans routinely lob a “you’re Orwellian” argument at their rivals, and this opinion goes there too. The dissent says:

Power must be tempered by accountability….On the one hand, platforms have developed the ability to monitor and control how all of us use the internet, exercising a power reminiscent of an Orwellian nightmare

The most Orwellian angles in this case are the dissenting judges’ abuse of their power without accountability and their ability to engage in judicial doublespeak.

It’s also bizarre and mockable to direct concerns about “unrestrained omnipotence” at…Snap? On this list of top apps, Snap ranks #17, right behind the Samsung Calendar. Snap isn’t exactly controlling “how all of us use the internet.” Indeed, I’m guessing Snap has “monitored” or “controlled” precisely zero of the dissenting judges. I’m even more confident I’ve escaped Snap’s omnipotence (never used it, don’t plan to).


The dissent ends with a plea to the Supreme Court: “it is once again up to our nation’s highest court to properly interpret the statutory language enacted by Congress in the Communications Decency Act.” I’m sure the 7-to-8 split and the Justice Thomas shoutouts will draw some attention from the justices. Is it enough to get 4 votes for certiorari?

Meanwhile, 7 Fifth Circuit judges signed onto this dissent, and other partisan judges (e.g., Judge Jones, who agreed with Judge Oldham’s #MAGA opinion in NetChoice v. Paxton, and well-known SCOTUS auditioner Judge Ho) didn’t sign but might still be open to revisiting Section 230. As a result, any future Section 230 case heading to the Fifth Circuit faces an extreme risk of judicial activism to overturn the existing Fifth Circuit precedent and disrupt decades of Section 230 jurisprudence. Something to look forward to.

Case Citation: Doe v. Snap, Inc., 2023 WL 8705665 (5th Cir. Dec. 18, 2023)