Adam Kovacevich has defined the “Prager Effect” as “suing Big Tech to make your MAGA censorship allegation, resulting in Courts significantly strengthening platforms’ legal rights to moderate.” Prager, which makes misleading videos that appear designed to radicalize kids to the MAGA agenda, sued YouTube over its demonetization decisions, claiming that YouTube had engaged in biased content moderation. That lawsuit produced a game-changer Ninth Circuit decision that has become the flagship federal opinion on account terminations and content removals.
While the Ninth Circuit case was pending, Prager doubled down by initiating a parallel state court lawsuit. Prager advanced four claims: California Constitution article 1, section 2 (free speech); Unruh Act violations; UCL; and breach of the implied covenant of good faith and fair dealing. YouTube defended on Section 230 and other grounds. See my blog post on the lower court ruling. The resulting state court lawsuit has reinforced the Prager Effect, producing another published opinion thoroughly rejecting every argument advanced by Prager and expanding services’ legal protections for their moderation decisions. The appellate court summarizes its conclusion:
However denominated, Prager’s state law causes of action target providers of an interactive computer service in their capacity as publishers restricting access to Prager’s information content—i.e., the same “Good Samaritan” screening that Congress has elected to protect from liability under state law. To the extent Prager has pleaded alternative theories of contractual breach and fraudulent business practices to forestall this conclusion, it has neither identified a contractual provision to support the former nor alleged facts establishing standing to prosecute the latter, and it raises no reasonable possibility of curing these defects by amendment. We further reject Prager’s argument that the CDA—foreclosing as it does Prager’s causes of action against private actors—violates the First and Fourteenth Amendments.
Section 230 applied to most of Prager’s claims. Citing Barrett v. Rosenthal, the court says:
section 230 protects an interactive computer service provider’s curation of content on its platform from “ ‘ “claims that would place a computer service provider in a publisher’s role. Thus, lawsuits seeking to hold a service provider liable for its exercise of a publisher’s traditional editorial functions—such as deciding whether to publish, withdraw, postpone or alter content—are barred.”…
Whether styled as a violation of the California Constitution’s guarantee of free speech and association, the Unruh Act’s antidiscrimination provisions, the UCL, or defendants’ terms of service, the conduct Prager alleges is injurious consists of defendants’ decisions regarding the audience to which videos would be published and whether publication of the videos would include paid advertising. Because Prager seeks to hold defendants liable under state law for these editorial publication decisions, section 230(c)(1) and (e)(3) forecloses relief
Note how this explicitly covers both leave-up and takedown decisions. Section 230 applies to all content moderation decisions.
The opinion then demolishes the standard grasping-at-straws “but the algorithms…!” argument:
Prager’s contention that defendants are themselves an information content provider—in that they developed algorithms used in determining whether to restrict access to Prager’s videos—does nothing to defeat section 230 immunity. Prager pleads no facts from which defendants’ use of algorithms would render them providers of information content. What Prager alleges is the use of “an automated filtering algorithm that examines certain ‘signals’ like the video’s metadata, title, and the language used in the video. The algorithm looks for certain ‘signals’ to determine if rules or criteria are violated so as to warrant segregation in Restricted Mode.” To the extent that an automated filtering algorithm is itself information, defendants of course created it; what is also apparent from Prager’s pleaded facts, however, is that defendants have not “provided [it] through the Internet or any other interactive computer service” within the meaning of section 230(f)(3), to Prager or anyone else…
Prager cites no authority for the proposition that algorithmic restriction of user content—squarely within the letter and spirit of section 230’s promotion of content moderation—should be subject to liability from which the algorithmic promotion of content inciting violence has been held immune…
Prager’s claims turn not on the creation of algorithms, but on the defendants’ curation of Prager’s information content irrespective of the means employed: it is not the algorithm but Prager’s content which defendants publish (or depublish). To the extent Prager’s claims principally rest on allegations that defendants violated a duty under state law to exercise their editorial control in a particular manner, defendants are immune under section 230 from the claims Prager brings in this suit.
In case you missed it: the court says, unreservedly, that “algorithmic restriction of user content” is “squarely within the letter and spirit of section 230’s promotion of content moderation.” This is worth reinforcing because the pro-censorship forces have lined up against this common-sense principle. Indeed, the whole “but the algorithms” attack on Section 230 has always been nonsensical. Sadly, I doubt this passage will stop its continued invocation.
Prager tried a variety of contract-based workarounds to Section 230. The court sets the baseline: “Section 230(c)(1) and (e)(3) do not necessarily foreclose contractual claims, however, where a computer service provider has agreed to limit its exercise of editorial discretion according to bargained-for terms and conditions…” However, evidence of this is rare–this is really just the Barnes promissory estoppel workaround, which rarely works in practice. Here, the problem is that YouTube’s contracts say it can do whatever it wants, and the attempts to form contract-like obligations from other public statements is tendentious and futile:
The Murphy court, and others, have held that the CDA foreclosed liability where plaintiffs have identified no enforceable promise allegedly breached…Prager’s contractual theories are barred because they are irreconcilable with the express terms of the integrated agreements….
the written contracts governing Prager’s relationship with defendants—limited to YouTube’s Terms of Service (YouTube TOS) and Google LLC’s AdSense Terms of Service (AdSense TOS), which the trial court judicially noticed without objection— contain no provision purporting to constrain defendants’ conduct as publishers…
Though consistent with Prager’s assertion that YouTube makes public-facing representations giving the impression that it voluntarily filters the content on its platform using a discrete set of neutral policies, the Community Guidelines in no way purport to bind defendants to publish any given video, or to remove a video only for violation of those guidelines….
As with the Community Guidelines, Prager conflates user guidelines with provider duties. Prager does not explain how defendants’ illustration in the guidelines of unsuitable content that “will result in a ‘limited or no ads’ monetization state” confers on users a contractual right that all other user content be monetized. At most, the Advertiser-friendly content guidelines permit users to “request human review of [monetization] decisions made by [defendants’] automated systems.” Thus, neither the Community Guidelines nor the Advertiser-friendly guidelines conflict with or limit defendants’ express reservation of rights….
the CDA may permit a state law claim concerning publishing activity based on a specific contractual promise, section 230 notwithstanding; this does not mean that the CDA requires an express contractual reservation of publishing discretion as condition precedent to section 230 immunity from state law claims…
Prager also tried a mockable argument that recently failed in Divino v. Google:
Prager suggests, however, that the California choice-of-law provision in YouTube’s Terms of Service (YouTube TOS) implies defendants’ tacit agreement to waive CDA immunity as to all claims arising under state law, rather than Prager’s agreement that California law would govern claims arising under the contract. Prager’s argument “misapprehends the meaning of the choice-of-law provision”
Seriously, this argument is so terrible that it would score zero points if a student tried it on one of my exams. But to MAGA plaintiffs, I guess the argument is so nice it’s worth failing twice.
In a footnote, the court rejects the plaintiff’s Hail Mary that the Fourth Circuit’s burn-down-230 opinion in Henderson somehow helped here:
Henderson’s narrow interpretation of section 230(c)(1) is in tension with the California Supreme Court’s broader view (see Murphy, supra, 60 Cal.App.5th at pp. 24-26 [applying Barrett and Hassell]), which we follow, absent a contrary ruling by the United States Supreme Court.
Does this start to suggest a real circuit split on 230? (Which was conspicuously absent in the Gonzalez case). Either way, expect anti-23o plaintiffs to invoke Henderson regularly. I hope they get the same resolution as Prager gets here.
Other Promise-Based Claims
The implied covenant of good faith and fair dealing claim fails even outside of 230:
Prager contends that the implied covenant of good faith and fair dealing obligated defendants to make publishing decisions in a manner Prager alleges good faith requires. Prager’s contractual claim is thus not for the enforcement of any express promise, but for imposition of a duty that would be at odds with the express reservation of defendants’ unfettered discretion in making publishing decisions
Going beyond the contracts, Prager looks to various “promises” it alleges that defendants made through public-facing comments. But, as Prager variously argues or concedes, the parties entered integrated contracts expressly providing that the written agreements, together with the materials incorporated by reference, constitute the “entire agreement” concerning each relevant service. Consistent with its concession that the parties entered integrated contracts, Prager has not articulated any basis for treating defendants’ public-facing comments as contractual…
defendants’ generalized public statements regarding their monitoring and filtering practices do not give rise to a state law contractual obligation to regulate their publishing decisions
A fraud argument fares no better:
In the trial court, Prager contended that its UCL claim was based primarily on defendants’ alleged misrepresentations regarding their operation YouTube as an open “forum for ‘freedom of expression’ for all,” their use Restricted Mode as a neutral content moderation tool, and the reasons that they restricted access to Prager’s videos. In its appellate briefing, Prager stated that the injuries it alleged in its complaint “were directly and proximately ‘caused by [defendants’] decisions to restrict and/or demonetize [Prager’s] content.’ ” Prager says that “the harm and injuries alleged flow directly from [defendants’] breaches of the Contracts, not ‘from [Prager’s] decision to use YouTube.’ ” In other words, Prager alleges no harm from what the Ninth Circuit has deemed defendants’ “lofty but vague statements” of “classic, non-actionable opinions or puffery,” only from defendants’ exercise of publishing discretion.
Prager tried an anemic attack on Section 230’s constitutionality. It goes nowhere:
State action is absent here, because social media platforms are generally permitted to decide for themselves what content to publish…state incursions into defendants’ discretion to restrict content on its platform may pose a more significant First Amendment concern than defendants’ restrictions…. [cites to NetChoice v. FLA]
Even assuming Prager were capable of amending its complaint to narrow its generic claims of “Protected Identity” to a classification the Fourteenth Amendment treats as suspect, it is Prager that claims a right to enforce what it would prefer to be the terms of the operative contract. This alone divests Prager of any legitimate means of styling itself as the victim of Jim Crow contract enforcement vis-à-vis defendants’ limited incursion on Prager’s ability to realize ad revenue from YouTube views….
[Eric’s comment: If you’re invoking Shelley v. Kramer to try to entrench your majoritarian position, haven’t you lost sight of everything?]
the thrust of Prager’s argument is that the government has a constitutional mandate to prevent private entities from discriminating on the basis of any protected classification and the failure to do so—that is, the trial court’s dismissal of a claim for relief from discrimination—is a state action in violation of that mandate. But the logical corollary of the requirement of state action is the existence and persistence of private discrimination that is beyond the reach of a self-executing constitutional right. Put differently, while Congress can surely enact legislation to prevent and remedy invidious discrimination (see, e.g., 42 U.S.C. § 2000e, et seq.), if it could be said that the Constitution requires Congress to enact antidiscrimination legislation in all private contexts, the state action requirement would be a nullity—the Constitution would effectively ban all discrimination by both public and private actors because state action would always be present. That is not the law.
It’s not much fun to mock Prager and its attorneys for weaponizing anti-discrimination laws to try to protect their majority status. After all, taking the plaintiff at face value, it seeks to impose government censorship to reduce Constitutionally protected editorial freedoms–with all of the implicit downsides I enumerate here. On the other hand, it is great fun to mock Prager and its attorneys for contributing to the Prager Effect. I guess we should thank them for their public service. ?
While Prager goes down in flames again, the war isn’t over. It’s now being fought in the Supreme Court, which will opine on an algorithmic recommendation exception to 230 in Gonzalez and will likely address the Florida and Texas cases involving statutory codifications of the rule of law Prager was litigating to install. The fate of the Internet hangs in the balance, and the partisan dimensions of these issues raise grave concerns about what the Supreme Court will do.
Case citation: Prager U. v. Google LLC, 2022 WL 17414495 (Cal. App. Ct. Dec. 5, 2022)