The Fifth Circuit’s Campaign to Undermine Section 230 Is Making Progress–AB v. Salesforce

This is a FOSTA case involving Backpage, but it’s against Backpage’s vendor Salesforce. That makes this a tertiary liability claim. Here are the steps in the plaintiffs’ alleged chain of liability:

V) victims promoted in Backpage advertisements (plaintiffs) –>

D1) advertisers (alleged sex traffickers) [1st party perpetrators] –>

D2) publishers (Backpage) [provided advertising support for perpetrator] –>

D3) vendor to publisher (Salesforce) [provided “back-office business services” to publisher, who provided advertising support to perpetrators who allegedly harmed the victims]

It should be obvious from this flowchart how far away the vendor is from the victim–and how little knowledge the vendor will have about any individual it’s never dealt with. Also, the publisher Backpage has many other vendors who provide it with “back-office business services” to support its publishing operations. Some or all of them become potentially become legally exposed if tertiary liability claims are tenable.

Tertiary liability isn’t automatically precluded by standard tort doctrines, but it puts a lot of stress on those doctrines. How can a “back-office business services” vendor owe a duty to its customer’s customer’s customer, especially when the plaintiffs concede that “Salesforce had ‘nothing to do’ with the trafficking advertisements posted on Backpage”? How can the “back-office business services” vendor be deemed the proximate cause of any harms with two other defendants in front of it?

Tertiary liability also creates significant evidentiary challenges. For the plaintiffs to win against Salesforce (the tertiary defendant), they will need to show that the primary and secondary defendants committed legal violations (i.e., there’s no tertiary liability if no one committed a direct violation). A tertiary defendant is in a terrible position evidence-wise to refute the liability of the upstream defendants. To the extent that courts have acknowledged this issue, they have presumed that everyone knew Backpage was providing illegal services to sex traffickers. So for purposes of a motion to dismiss, courts can simply assume the upstream violations. However…no court held Backpage liable for sex trafficking before it was shut down–despite many, many legal attempts to impose such liability. As a result of this implicit clean bill of legal health from the courts, vendors to Backpage had no independent reason to conclude that Backpage was engaging in illegal activities at the time they worked together.

[Note: Backpage shut down April 2018, but Backpage-related litigation continues into 2025 and beyond].

Add Section 230 into the mix and tertiary liability looks even weirder. Backpage routinely qualified for Section 230 protection in court (whether it still could do so isn’t addressed in this opinion). Could Backpage’s vendors be liable for the primary perpetrator’s actions if Backpage itself isn’t liable for its (more proximate) actions? Jumping over an immunized secondary defendant to impose liability against a tertiary defendant would eviscerate Section 230. Plaintiffs could simply avoid Section 230 by suing a publisher’s vendors. If so, many vendors will be pressured not to accept lawful engagements with paying customers because they would face impossible-to-manage potential liability exposure.

* * *

Despite the serious problems with tertiary liability in Section 230 cases, the courts have entertained it. The flagship case in this area is G.G. v. Salesforce, a 7th Circuit opinion from 2023 (which I ran out of time to blog) that embraced tertiary liability to hold that Section 230 didn’t apply to Salesforce’s services to Backpage.

The latest ruling comes from the Fifth Circuit, which reaches the same conclusion regarding Salesforce’s eligibility for Section 230 for providing vendor services to Backpage. In light of the GG precedent, it’s not a surprise. Still, every case ratifying tertiary liability is a troubling one, and this is no exception.

The ruling also fits into the broader Section 230 jurisprudence in the Fifth Circuit, which has swung decidedly against it. In Doe v. Snap, the Fifth Circuit came within one vote of reviewing Section 230 en banc so it could be gutted, and the Paxton v. NetChoice case delivered a mangled interpretation of Section 230 even though Section 230 wasn’t the core issue in that case (the First Amendment was). In light of the strong hostility to Section 23o among the Fifth Circuit judges, every Section 230 case that reaches the Fifth Circuit is likely to be a jurisprudential roller-coaster.

[Note: the opinion was written by Trump 1.0-appointed Judge Morales from the Southern District of Texas, sitting by designation. I’m always a bit puzzled when circuit law is being written by someone who, you know, isn’t actually a circuit judge. I raised this issue, for example, with the Ninth Circuit’s YOLO opinion, authored by a visiting 6th Circuit judge. Then again, maybe if an actual Fifth Circuit judge had written this opinion, it would have displayed even more antipathy towards Section 230?]

* * *

Salesforce argued that the plaintiffs’ claims treat it as the publisher or speaker of third-party content. The appeals court disagrees:

The text of section 230, our precedent, and the precedent of our sister circuits uniformly reject the argument Salesforce advances.

The court lays out its view of tertiary liability and says that Section 230 doesn’t automatically apply to a downstream defendant, even if an upstream defendant would be immunized by Section 230 because the claim involves publishing or speaking:

To the extent that our cases since MySpace have understood the section-230 analysis as focusing on the specific claims and allegations advanced by a plaintiff, those cases have remained faithful to the text of the statute Congress enacted…

In contrast to our claims-analysis approach, we have rejected a mechanical, but-for reading of section 230. A but-for test that asks whether third-party speech lies anywhere in the chain of causation leading to the alleged harm would expand section-230 immunity beyond the statute’s text. Such a test would align more with a statute that read “shall be held liable for conduct involving third-party speech.” But that is not the statute Congress enacted….

the emphasis, properly placed, would emphasize whether claims treat a defendant as a publisher or speaker, not whether third-party speech lies somewhere in the chain of causation…

other federal courts have rejected a mechanical approach to the publisher-or-speaker question. In MySpace, we held that a plaintiff cannot artfully plead around section 230 by casting their claims in terms of first-party conduct when they really seek to enforce a duty traditionally associated with publication. In other words, it does not matter whether the plaintiff styles the claims in terms of publication. It similarly does not matter whether publication is an explicit element of the plaintiff’s claim. And critically, like we held in Paxton, the fact that third-party speech is involved somewhere in the chain of causation that led to a plaintiff’s injuries does not mean that a plaintiff’s claims necessarily treat a defendant as a publisher or speaker of that third-party speech.

The court then says Salesforce doesn’t automatically qualify for Section 230:

According to Salesforce, because the only link between Salesforce and Plaintiffs’ sex-trafficking-related harms is the illicit ads their traffickers posted on Backpage, Plaintiffs necessarily seek to treat Salesforce as the publisher or speaker of those ads.

Although some daylight may peek through between Salesforce’s theory of section 230—let’s call it the only-link theory—and a but-for analysis, we close the blinds on its theory for two reasons.

The first reason the court closes the blinders: “Under this theory, any time third-party content is the only link in the chain of causation, section 230 provides immunity—encompassing vastly more than claims that seek to treat a defendant as a publisher or speaker of third-party content.”

The second reason the court closes the blinders: Something something about avoidable outcome anomalies. I didn’t understand the court’s second point. Either way, the court is tone-deaf if it complains about anomalies and doesn’t address the elephant in the room: that the tertiary defendant may face more liability than the secondary defendant, even the only reason any defendant is liable is the publication of third-party content.

The court says the plaintiffs’ allegations were not publisher/speaker-based, so they get around Section 230:

Plaintiffs allege that Salesforce knowingly assisted, supported, and facilitated sex trafficking by selling its tools and operational support to Backpage even though it knew (or should have known) that Backpage was under investigation for facilitating sex trafficking. In essence, Plaintiffs allege that Salesforce breached a statutory duty to not knowingly benefit from participation in a sex-trafficking venture.

To state the obvious: this duty does not derive from Salesforce’s status or conduct as a publisher or speaker and would not require Salesforce to exercise publication or editorial functions to avoid liability. Rather, the duty simply requires that Salesforce not sell its tools and operational support to a company it knew (or should have known) was engaged in sex trafficking.

Hold on. In the last sentence, the court uses the generic descriptor “a company.” That’s deceptively imprecise. Salesforce didn’t just serve any “company.” It served Backpage, which is liable (if at all) for publishing third-party ads. If the sentence is rewritten to be more precise, the court’s reasoning looks a lot more suspect.

Having failed with the should-be-winning argument that Salesforce’s liability only derives from the publication of third-party ads, Salesforce tried a secondary argument that its liability is based on “publishing” Backpage’s (third-party) data back to Backpage. The court says the Seventh Circuit rejected that argument in the GG ruling.

The court also says that “While Salesforce argues it would be required to police the use of its products, that is not quintessentially the function of a publisher.” I mean…what is the quintessential function of a “publisher” if it’s not to moderate content and exercise editorial control over it? The court responds:

a publisher might also need to police the use of its products. But that is not because of the publisher’s unique functions. Rather, it is because a publisher—like Salesforce and every other entity subject to United States federal law—owes a statutory duty to the public not to knowingly benefit from participation in a sex-trafficking venture.

If the statutory duty is violated by “publishing” content, then the claim still treats Salesforce as a “publisher.” Otherwise, you could say that every claim against a publisher stems from a non-publisher “duty.” (E.g.: Defamation? Comes from a duty not to communicate false statement of facts). In other words, by interpreting the “knowingly benefit from participation” as a different duty than “publishing,” when the defendant’s only activity is publishing content, the court is playing word games.

This discussion illustrates why the “duty” analysis articulated in Calise and YOLO did serious doctrinal damage to Section 230. It encourages this kind of semantic hair-splitting that suggest that the “duty” at issue differs from publishing, even when they are really the same thing. Historically, courts have gotten Section 230 rulings right by treating ALL claims as publisher/speaker claims when the claims are based on third-party content. The “duty” analysis pushes courts to distinguish between the plaintiff’s claims and the functions of publishers.

I think the court concedes that its distinction is illusory with this acknowledgement:

To be sure, a plaintiff might use section 1595 or section 98.002 to artfully plead around section-230 immunity. For example, a plaintiff might allege that a defendant knowingly benefitted from a sex-trafficking venture by failing to implement adequate content-moderation policies. Although pled in terms of first-party conduct, this claim would ultimately seek liability “for decisions relating to the monitoring, screening, and deletion of content from its network—actions quintessentially related to a publisher’s role.” In such a case, section-230 immunity would likely attach

This is exactly what the plaintiffs are doing: saying that Salesforce is liable because it didn’t adequately content-moderate Backpage’s data. So when the court says “Plaintiffs seek to hold Salesforce liable for allegedly providing back-office business services to a company it knew (or should have known) was engaged in sex trafficking,” the plaintiffs have gotten the judge to accept their synonym for publishing (“back-office business services”) as something different.

To reinforce that it’s splitting hairs, the court purports to enumerate all of the things Salesforce allegedly isn’t doing:

The evidence shows that Salesforce did not have any role in:

• screening, monitoring, or filtering content;

• reviewing or analyzing third-party content;

• transmitting or hosting third-party content;

• editing or altering third-party content;

• developing or enforcing content-moderation policies; or

• deciding how third-party content was organized or displayed

This is pure advocacy. Given that Salesforce is arguing that Backpage’s data is the “third-party content” at issue (in this secondary argument), the plaintiffs are in fact trying to hold Salesforce liable for doing/not doing all of these things.

In a footnote, the court says without further comment that “Plaintiffs agree that Salesforce had ‘nothing to do’ with the trafficking advertisements posted on Backpage.” That seems like a critical concession, no? If Salesforce had “nothing to do” with the ads that caused the harm, then how could it be liable for those ads? The court’s answer appears to be that Salesforce never should have become a Backpage vendor at all, even though Backpage was never declared to be engaged in illegal services during its lifespan.

* * *

Finally…as I’ve written before:

I am incapable of talking about Salesforce being denied Section 230 immunity without snarkily reminding you that Salesforce’s CEO, Benioff, myopically called for the repeal of Section 230. Still hate 230, Benioff?

Each time Salesforce loses its Section 230’s defense, perhaps Benioff’s anti-Section 230 position makes more sense. There’s not much downside to his self-interest in trashing Section 230 if you don’t qualify for it… But if Salesforce’s CEO truly believes in repealing Section 230, he could make a voluntary contribution to the effort by having Salesforce not claim it deserves Section 230 as a defense. That would have the side benefit of not horking Section 230 jurisprudence. Or perhaps Salesforce wants courts to scale back Section 230 as a common law repeal of the statute? So many ways for Benioff to win by losing.

Case Citation: A.B. v. Salesforce, Inc., 2024 WL 5163222 (5th Cir. Dec. 19, 2024)

Prior Blog Posts About Salesforce

* Five Decisions Illustrate How Section 230 Is Fading Fast
* The 7th Circuit’s Section 230 Jurisprudence’s Impact on FOSTA Cases
Section 230 Helps Salesforce Defeat Sex Trafficking Lawsuit–GG v. Salesforce
* Catching Up on Recent FOSTA Developments (None of Them Good)
Section 230 Helps Salesforce Defeat Sex Trafficking Lawsuit–Doe v. Salesforce

More SESTA/FOSTA-Related Posts

* Omegle Defeats Lawsuit Over User’s “Capping”–MH v. Omegle
Section 230 Immunizes OnlyFans for User-Uploaded Video–Doe v. Fenix
Five Decisions Illustrate How Section 230 Is Fading Fast
Section 230 Preempts FOSTA Claim–Doe v. WebGroup Czech Republic
Instagram Defeats Lawsuit Claiming It Was a “Breeding Ground” for Sex Traffickers–Doe v. Backpage
The 7th Circuit’s Section 230 Jurisprudence’s Impact on FOSTA Cases
Grindr Defeats FOSTA Claim–Doe v. Grindr
Twitter Defeats FOSTA Case Over CSAM–Doe v. Twitter
DC Circuit Upholds FOSTA’s Constitutionality (By Narrowing It)–Woodhull v. U.S.
Section 230 Immunizes Snap, Even if It’s “Inherently Dangerous”–L.W. v. Snap
The Ninth Circuit’s FOSTA Jurisprudence Is Getting Clearer (and More Defense-Favorable)
Defendants Get Important FOSTA Win in 9th Circuit–Doe v. Reddit
More Evidence that FOSTA Benefited No One
Omegle Denied Section 230 Dismissal–AM v. Omegle
Section 230 Helps Craigslist Defeat Sex Trafficking Case–LH v. Marriott
Section 230 Helps Salesforce Defeat Sex Trafficking Lawsuit–GG v. Salesforce
Constitutional Challenge to FOSTA Fails–Woodhull v. US
Catching Up on a FOSTA Case–ML v. Craigslist
Facebook Loses Jurisdictional Ruling in Texas Sex Trafficking Lawsuit–Facebook v. Doe
Justice Thomas Really, REALLY Wants Section 230 Repealed (Even If He Has to Do It Himself)
Section 230 Immunizes TikTok for User-Posted Videos–Day v. TikTok
So Many Unanswered Empirical Questions About FOSTA
Another Problematic FOSTA Ruling–Doe v. Pornhub
Catching Up on Recent FOSTA Developments (None of Them Good)
Section 230 Preempts Claims Against Omegle–M.H. v. Omegle
To No One’s Surprise, FOSTA Is Confounding Judges–J.B. v. G6
FOSTA Claim Can Proceed Against Twitter–Doe v. Twitter
FOSTA Survives Constitutional Challenge–US v. Martono
2H 2020 Quick Links, Part 4 (FOSTA)
Justice Thomas’ Anti-Section 230 Statement Doesn’t Support Reconsideration–JB v. Craigslist
Sex Trafficking Lawsuit Against Craigslist Moves Forward–ML v. Craigslist
Section 230 Preempts Another FOSTA Claim–Doe v. Kik
Section 230 Protects Craigslist from Sex Trafficking Claims, Despite FOSTA–JB v. Craigslist
Facebook Still Can’t Dismiss Sex Trafficking Victims’ Lawsuit in Texas State Court
Craigslist Denied Section 230 Immunity for Classified Ads from 2008–ML v. Craigslist
2H 2019 and Q1 2020 Quick Links, Part 3 (FOSTA/Backpage)
New Paper Explains How FOSTA Devastated Male Sex Workers
FOSTA Constitutional Challenge Revived–Woodhull Freedom Foundation v. US
New Civil FOSTA Lawsuits Push Expansive Legal Theories Against Unexpected Defendants (Guest Blog Post)
Section 230 Helps Salesforce Defeat Sex Trafficking Lawsuit–Doe v. Salesforce
Latest Linkwrap on FOSTA’s Aftermath
Section 230 Doesn’t End Lawsuit Claiming Facebook Facilitated Sex Trafficking–Doe v. Facebook
New Essay: The Complicated Story of FOSTA and Section 230
Who Benefited from FOSTA? (Spoiler: Probably No One)
FOSTA’s Political Curse
FOSTA Doesn’t Help Pro Se Litigant’s Defamation Claim Against Facebook
Constitutional Challenge to FOSTA Dismissed for Lack of Standing (Guest Blog Post)
An Update on the Constitutional Court Challenge to FOSTA–Woodhull Freedom v. US (Guest Blog Post)
Indianapolis Police Have Been “Blinded Lately Because They Shut Backpage Down”
Constitutional Challenge Against FOSTA Filed–Woodhull v. US (Guest Blog Post)
Catching Up on FOSTA Since Its Enactment (A Linkwrap)
More Aftermath from the ‘Worst of Both Worlds FOSTA’
‘Worst of Both Worlds’ FOSTA Signed Into Law, Completing Section 230’s Evisceration
Backpage Loses Another Section 230 Motion (Again Without SESTA/FOSTA)–Florida Abolitionists v. Backpage
District Court Ruling Highlights Congress’ Hastiness To Pass ‘Worst of Both Worlds FOSTA’– Doe 1 v. Backpage
More on the Unconstitutional Retroactivity of ‘Worst of Both Worlds FOSTA’ (Guest Blog Post)
Senate Passes ‘Worst of Both Worlds FOSTA’ (Linkwrap)
Why FOSTA’s Restriction on Prostitution Promotion Violates the First Amendment (Guest Blog Post)
SESTA’s Sponsors Still Don’t Understand Section 230 (As They Are About to Eviscerate It)
Can the ‘Worst of Both Worlds FOSTA’ Be Salvaged? Perhaps…and You Can Help (URGENT CALL TO ACTION)
Congress Probably Will Ruin Section 230 This Week (SESTA/FOSTA Updates)
What’s New With SESTA/FOSTA (January 17, 2018 edition)
New House Bill (Substitute FOSTA) Has More Promising Approach to Regulating Online Sex Trafficking
* My testimony at the House Energy & Commerce Committee: Balancing Section 230 and Anti-Sex Trafficking Initiatives
How SESTA Undermines Section 230’s Good Samaritan Provisions
Manager’s Amendment for SESTA Slightly Improves a Still-Terrible Bill
Another Human Trafficking Expert Raises Concerns About SESTA (Guest Blog Post)
Another SESTA Linkwrap (Week of October 30)
Recent SESTA Developments (A Linkwrap)
Section 230’s Applicability to ‘Inconsistent’ State Laws (Guest Blog Post)
An Overview of Congress’ Pending Legislation on Sex Trafficking (Guest Blog Post)
The DOJ’s Busts of MyRedbook & Rentboy Show How Backpage Might Be Prosecuted (Guest Blog Post)
Problems With SESTA’s Retroactivity Provision (Guest Blog Post)
My Senate Testimony on SESTA + SESTA Hearing Linkwrap
Debunking Some Myths About Section 230 and Sex Trafficking (Guest Blog Post)
Congress Is About To Ruin Its Online Free Speech Masterpiece (Cross-Post)
Backpage Executives Must Face Money Laundering Charges Despite Section 230–People v. Ferrer
How Section 230 Helps Sex Trafficking Victims (and SESTA Would Hurt Them) (guest blog post)
Sen. Portman Says SESTA Doesn’t Affect the Good Samaritan Defense. He’s Wrong
Senate’s “Stop Enabling Sex Traffickers Act of 2017”–and Section 230’s Imminent Evisceration
The “Allow States and Victims to Fight Online Sex Trafficking Act of 2017” Bill Would Be Bad News for Section 230
WARNING: Draft “No Immunity for Sex Traffickers Online Act” Bill Poses Major Threat to Section 230
The Implications of Excluding State Crimes from 47 U.S.C. § 230’s Immunity