Rose v. SpaceXAI LLC, 2026 WL 3169664 (N.D. Cal. Oct. 22, 2026)

Some consumers claim X employs ““deceptive and illegal ‘automatic renewal’ tactics to trick consumers” into signing up for “unwanted, pricey subscriptions” to xAI’s SuperGrok chatbot.” (As I observed, anyone paying to use Grok may be opting-into caveat emptor territory). X invoked the arbitration clause in its TOS. The court denies arbitration.

The court calls this screen a sign-in-wrap:

The court says the TOS offer isn’t reasonably conspicuous.

Transaction Context

Rose created a free Grok account in March 2025 through xAI’s sign-up page that invited users to “Create your account.” Rose created an account by clicking a button below the “Create your account” message that said, “Sign up with email.” Rose does not dispute that this sign-up process for a free account, allowing Rose to easily interact with Grok in the future, indicates a “forward-looking relationship” with Grok rather than a single transaction. A user signing up for a Grok account, like Rose, thus would expect the relationship to be governed by terms.

Visual Display

Despite Rose’s purportedly heightened interest in finding the TOS offer, the court says this is basically Skaggs redux.

Rose signed up for a Grok account using his email address—the second button down in the list of four sign up options. As in Skaggs, after Rose selected the “Sign up with email” button, the remaining text on the page became superfluous; Rose had no reason to continue reading. The notice of the “Terms of Service” at the very bottom of the screen was not sufficiently conspicuous during Rose’s signup process…

an “uncluttered” page is more likely to focus the reader’s attention on a notice of terms of service even where the text of the notice is muted, as here. But the notices analyzed in the controlling Ninth Circuit cases (Keebaugh, Massel, and Patrick) that Defendants cite were both uncluttered and required users to explicitly acknowledge the terms—a requirement xAI chose not to include in its signup process here. [Eric’s comment: the cited precedent isn’t as decisive as the court suggests.]…

xAI permits a user to create an account simply by clicking one of the sign-up options without verifying that the user is aware of—much less agrees to be bound by—the terms of service purportedly governing the relationship.

In a footnote, the court distinguishes another precedent: “here, the disclosure is at the periphery of the screen: a user signing up from a computer screen must survey a larger area to find the disclosure, and is thus less likely to see it, than a user signing up from a mobile device—even if the sign-up screen is uncluttered.”

The court doesn’t expressly address the mismatch between the TOS offer language “by continuing” and the action button labels “sign up with…” but that would be another potentially fatal flaw per Chabolla and Godun.

Chen v. X Corp., 2026 WL 3082938 (S.D.N.Y. Oct. 15, 2026)

This is a pro se account termination lawsuit, so it has no chance of success. Chen sued X in New York. X seeks to transfer it to Texas per its TOS. The court agrees.

The court doesn’t show any screenshots but says:

When Chen created his X Account, a blue underlined hyperlink containing X Corp.’s Terms was located near the “Create My Account” button. The signup page also noted that in creating an account, a user agreed to the Terms. The Terms included a clause that X Corp. may change the Terms and that continued use of the platform after notice of the change in the Terms would constitute acceptance of the Terms. Further, on October 21, 2024, Chen clicked “Got It” on a pop-up notification notifying users that X Corp. was updating its Terms. The relevant forum selection clause was contained in these updated Terms. Chen does not dispute that there was a hyperlink to X Corp.’s Terms when he signed up, nor that he clicked “Got It” on October 21, 2024.

Query: is “Got it” the same as an acceptance of an offer? In my Internet Law class, I show students this screenshot and mock it because it has soft language (“now’s a great chance”–is it really though?) and nothing in the disclosure says that the “got it” button signals assent.

The court says:

The Terms were reasonably communicated to Chen through the pop-up notification he received on October 21, 2024, informing him that the Terms were being updated. This Circuit has also frequently held that terms communicated in a sign-up wrap format similar to X Corp.’s satisfy the reasonably conspicuous notice requirement.

In a pro se case, X can get away with arguments like this. Had a competent lawyer been challenging this screen in front of an impartial judge, I’m not so sure. Then again…

X.AI LLC v. St. Clair, 7:26-cv-00005-O (N.D. Tex. Oct. 20, 2026)

This is part of a litigation brawl between Elon Musk and Ashley St. Clair, one of Musk’s baby mamas who has evolved into a fierce Musk critic. St. Clair sought a TRO in New York “alleging that Grok generated explicit images depicting [her] and propagated those images on the X platform.” Team Musk countersued St. Clair in Texas, invoking X’s TOS due to St. Clair having created an X account in 2024. Judge Reed O’Connor gives Team Musk everything they want, despite (or because of?) there being questions about his impartiality. Friendly judges are a key reason why Musk changed X’s venue selection to Northern District of Texas. He’s apparently getting good ROI on that switch.

This takes us back to the “got it” screen I just discussed in the Chen case. Once again it works, but this time in a hotly contested lawsuit. Judge O’Connor gives a clean sweep to Team Musk:

Plaintiffs assert that Defendant received notice of Version 21 on December 16, 2025, clicked the accompanying “Got It” button, and continued accessing the X platform after the revised terms became effective on January 15, 2026.22 Defendant argues that clicking the notification before the effective date could not constitute acceptance and that Plaintiffs’ allegations of subsequent use are conclusory. But Plaintiffs expressly allege that Defendant accessed or used X after January 15, 2026, conduct that the Revised Terms of Service identify as constituting acceptance….

Plaintiffs allege that Version 21 expressly extends the forum-selection provision to third-party beneficiaries and that the xAI Plaintiffs qualify as such beneficiaries. Although Defendant disputes whether the parties intended to include xAI within that provision, the absence of xAI’s name from the Terms of Service does not, standing alone, establish that Plaintiff’s interpretation is implausible….

Defendant maintains that her claims arise independently of any contractual relationship because they concern xAI’s alleged misconduct in generating sexually explicit images without her consent. But the forum-selection clauses are not expressly limited to contractual causes of action. Rather, they encompass disputes related to the services and content generated or disseminated through those services….Plaintiffs allege that Defendant’s New York action concerns content generated by Grok and disseminated through the X platform. Given the clauses’ express references to disputes concerning content and others’ use of the services, Plaintiffs have plausibly alleged that Defendant’s claims fall within their scope.

I’m still skeptical about the enforceability of that “Got It” button!

The post Catching Up on X’s TOS Formation Cases appeared first on Technology & Marketing Law Blog.

* * *

This is an account termination/content removal case, so you already know the plaintiff is going to lose–the only question is how.

Shared allegedly “creates and publishes original, timely, and entertaining [online] content.” It ran Facebook pages that, at one point, had 25M followers. Shared heavily used Facebook’s self-service advertising, placing over 500,000 ads. Facebook rejected over 18,000 of those ads. The court says “Shared regularly relied on ‘scammy’ ads and ‘blatant clickbait.'” Here is an example of one of their rejected ads:

This ad showed Luke Wilson next to a serious car crash, baiting viewers to click on the ad to check whether he was OK. (He was). Facebook rejected this ad based on its policy against “shocking or sensational content,” including photos that “depict violence.” Shared was upset that it got canned explanations for its ad rejections, rather than personalized roadmaps to correct any problems.

Ultimately, Facebook pulled the plug on Shared’s operations in October 2020. The court summarizes Shared’s claims:

Shared sued Meta for, among other things, conversion and fraudulent inducement, claiming that Meta’s decision to “unpublish[] the Shared [Facebook] Pages” was a business “death sentence.” It based the conversion claim on Meta’s “unpublishing” the Shared Facebook Pages, over which Shared argued it had an exclusive property right.

If none of this makes sense to you, join the club. Conversion is a chattel-based concept, so what is it doing here? It brought to mind the Best Carpet Values v. Google case, where the plaintiff unsuccessfully claimed that web page framing was a trespass to the chattels of intangible web pages. What is it with plaintiffs importing offline chattels-based doctrines into the intangible web?

The lower court dismissed the case. My prior blog post.

* * *

In a published opinion, the Ninth Circuit dismisses the conversion claim on Section 230 grounds. “Shared’s conversion claim seeks to impose liability on Meta for “unpublishing” Shared’s Facebook pages.” Labeling the claim as “conversion” was unusual but not creative or productive. The court responds flatly: “Section 230 bars the conversion claim.”

Shared only contested the publisher/speaker element, arguing that the conversion claim was somehow not that. The court easily rejects that:

Shared sued Meta because the latter exercised its publishing prerogative by removing third-party content—i.e., Shared’s Facebook pages. Such a decision to remove content and prevent someone from posting advertisements is a quintessential editorial decision protected under Section 230.

Shared tried to end-run this obvious conclusion by invoking the Calise/YOLO line of promise-based Section 230 exceptions:

Shared contends that its conversion claim effectively “seeks to hold Meta liable for failing to comply with its promise” not to remove Shared from Facebook “unless Shared ‘clearly, seriously, or repeatedly’ breached Facebook’s Terms or Policies,” which Shared claims it did not.

If you read Calise and YOLO literally, Section 230 should fall away once the plaintiff cites any promise made by the defendant. That did not happen here.

First, the court expresses skepticism about whether any promise was made because the TOS “gives Meta unilateral authority to determine whether a business has breached the Community Standards.” However, that’s not a 230 defense, that would be a failure of the prima facie elements. I’ll note that plaintiffs routinely mine a defendant’s site disclosures to identify and overstate alleged promises that are anything but.

Second,

Calise is of little help because Shared still seeks to meddle with Meta’s publishing discretion….a contractual promise that reiterates the company’s inherent publishing authority still amounts to publishing activity protected by Section 230. [Cite to King v. Facebook.] Indeed, it makes little sense to strip Meta of Section 230 liability defense simply because Facebook articulated its discretion as a publisher in its Terms of Service.

The court distinguishes Barnes: “Yahoo’s one-time promise to a specific user was beyond its typical publishing activity.” The court distinguishes Calise: “Facebook in Calise promised to curb advertisements that ‘defraud[] or exploit[]’ for ‘money or property.’ Such a contractual promise does not sound in editorial discretion but rather represents a business commitment to smoke out scammers.” Do those distinctions persuade you?

The court summarizes:

a claim does not automatically fall outside the ambit of Section 230’s liability protection just because it is framed as a breach of contract. Instead, a court must evaluate a claim to determine whether it seeks to treat an interactive computer service as a publisher or speaker of third-party information. That task may include assessing the core theory of the complaint and the nature of the alleged injurious conduct. In our case, Shared’s claim hinges on Meta’s publishing authority, despite being dressed up in contractual garb. Section 230 thus bars Shared’s claim.

In other words, Shared didn’t like having its account yanked, especially after all of the money it spent on Facebook ads, but it’s Facebook’s editorial prerogative to publish or not publish Shared’s content and ads. The net effect is straightforward: an account termination/content removal case fails even when a plaintiff tries the obtuse framing of conversion. No surprises there.

Yet, if you asked me to explain why Calise’s claim didn’t turn on Facebook’s editorial prerogative but Shared’s claim did, I couldn’t do it persuasively. Maybe one way to reconcile it is that Shared was publishing its own content on Facebook, while the plaintiffs in Barnes, Calise, and YOLO were suing based on someone else’s published content? That sure seems like a fine distinction, and the court didn’t make it clearly.  I imagine the Ninth Circuit will have plenty more opportunities to plumb its purported distinctions. I personally think this decision casts more doubt on the wisdom of both Calise and YOLO.

* * *

In a separate unpublished opinion, the court rejects Shared’s misrepresentation claim on statute of limitations grounds. “Shared received thousands of ad-rejection messages with allegedly inadequate explanations starting in 2016. But it did not sue until six years later.” But also, “even if Shared’s fraud claims were timely, Shared waived its fraud claims through its course of conduct over multiple years” because it kept placing orders despite receiving allegedly defective ad rejections.

Case Citation: The published opinion: Shared Partnership v. Meta Platforms, Inc., 2026 WL 3266791 (9th Cir. Oct. 30, 2026)

The unpublished opinion: Shared Partnership v. Meta Platforms, Inc., No. 25-2612 (9th Cir. Oct. 30, 2026)

Plaintiff’s lawyers were: Cook Alciati of Gardella Alciati, whose bio says he “specializes in intellectual property litigation with a particular emphasis on patent litigation,” and Seth Wiener of San Ramon.

The post Section 230 Blocks Lawsuits That “Meddle” With Online Publishers’ Editorial Decisions–Shared v. Meta appeared first on Technology & Marketing Law Blog.

Generated by ChatGPT Dec. 2025

Alex Berenson is a one-time N.Y. Times reporter whose life appears to have taken an unfortunate turn in the wake of MAGA and COVID. According to his Wikipedia entry:

During the coronavirus pandemic, Berenson appeared frequently in American right-wing media, spreading false claims about COVID-19 and its vaccines. He spent much of the pandemic arguing that its seriousness was overblown. Once the COVID-19 vaccines became available, he spread misinformation about the safety and effectiveness of the vaccines.

(A reminder that 1.2M Americans died due to COVID).

Berenson aggregated over 200k followers at Twitter who were exposed to his hot takes on COVID and vaccines. That put his account at risk as Twitter cracked down on misinformation and denialism:

In May 2020, Twitter announced that it may append labels or warning messages to tweets related to COVID-19 misinformation. Berenson tweeted that he was “concern[ed] about what the new rules might mean for free speech on [Twitter].” In response, Twitter’s Vice President of Global Communications Brandon Borrman contacted Berenson to discuss the new policy. In December 2020, Twitter announced a new policy with respect to COVID-19 vaccine misinformation, stating that it would start to “remove Tweets which advance harmful or misleading narratives about COVID-19 vaccinations.” Berenson contacted Borrman, “looking for assurances that Twitter would continue to allow his reporting on COVID-19 vaccines.” Borrman allegedly responded that the “points [Berenson was] raising should not be an issue at all,” and that Berenson should reach out if he “r[an] into any issues.”

On March 1, 2021, Twitter announced a “five-strike policy” to combat COVID-19 misinformation. Accounts that received strikes were subject to increasing levels of discipline: the first strike resulted in a warning, the second and third strikes each in a 12-hour account lock, the fourth strike in a 7-day account lock, and the fifth or successive strike in permanent suspension. The policy provided that, if a user “believe[s] that [their] account was locked or suspended in error, [the user] can submit an appeal.”

The next day, Berenson spoke with Borrman, who allegedly reported that “[Berenson’s] name has never come up in the discussions around these policies,” and that Borrman would “try to ensure [that Berenson would be] given a heads up before an action is taken, but [that Borrman was] not always made aware of them before they’re executed.”…

Berenson received strikes from Twitter in March, July, and August 2021. His Twitter account was permanently suspended with the fifth strike on August 28, 2021.

Berenson “alleges that former President Joseph R. Biden, as well as other individuals in the Biden Administration, conspired with executives at Pfizer, Inc. to pressure Twitter to suspend Berenson’s Twitter account, because of his views with respect to the COVID-19 vaccines.” He sued various US government officials for roles in the alleged censorship. In 2025, the district court dismissed several of the defendants. The Second Circuit, in a memo opinion, affirms that ruling.

Section 1985(3)

Berenson asserts that the relevant class here is a “class of Americans who had chosen not to receive a COVID-19 vaccine, a class of which [ ] Berenson is apart [sic].” The district court correctly held that Berenson’s proposed class does not constitute a class for Section 1985(3) purposes. Instead, this is precisely the kind of class that the Supreme Court cautioned against in Bray—a class composed of individuals who share a desire to engage in conduct, i.e., to forgo COVID-19 vaccination, that the defendants allegedly disfavor. To the extent Berenson suggests that the class falls within the ambit of Section 1985 because it “includes a disproportionate number of African-Americans, political conservatives, and evangelical Christians,” we are unpersuaded….

the FAC is replete with allegations that Berenson, and not a general class of unvaccinated Americans, was specifically targeted….Berenson’s Section 1985 claim must be dismissed because he failed to plausibly allege “that the defendants acted with class-based, invidiously discriminatory animus.”

Tortious Interference

Twitter’s five-strike policy and Borrman’s alleged remarks, whether considered separately or together, are not “wholly incompatible” with the Terms of Service. The relevant provision of the Terms of Service is located in the section entitled “Ending These Terms,” and provides the means by which a user may terminate the agreement…the provision makes clear that suspension or termination of an account is not limited to violations of that policy….

[Borrman’s] remarks suggest at most that Borrman would attempt to notify Berenson of action taken against him—not that no action would be taken against Berenson or that Berenson would necessarily receive advance notice of any action.

In 2022, Berenson’s deplatforming claim against Twitter partly survived a motion to dismiss. My blog post on that ruling. That ruling became a bit of a cause celebre in MAGA circles as proof that (1) Berenson was censored, and (2) Twitter could be held liabile. The Second Circuit rejects that ruling:

Berenson relies heavily on Berenson v. Twitter, Inc., No. C 21-09818 (WHA), 2022 WL 1289049 (N.D. Cal. Apr. 29, 2022), a decision in a case Berenson initiated against Twitter in December 2021. There, the district court granted in part and denied in part Twitter’s motion to dismiss, holding that the five-strike policy and Borrman’s “specific and direct assurances to [Berenson] regarding his posts pursuant to that policy” “amount[ed] to a modification of the [Terms of Service]” and that Berenson had plausibly alleged that “Twitter’s conduct … modified its contract with [him] and then breached that contract by failing to abide by its own five-strike policy and its specific commitments set forth through [Borrman].” We find that analysis unpersuasive

It’s a reminder that plaintiffs have virtually no chance of successfully suing Internet services for their account termination/content removal decisions. Even one of the few preliminary “wins” by plaintiffs is “unpersuasive” when reviewed by an (unrelated) appellate court.

The litigation over Biden-era COVID-related “censorship” feels quaint in light of the ongoing, relentless, and far more egregious censorship taking place daily in the Trump 2.0 administration. If anyone still wants to make censorship their cause celebre, it would make sense to start with the censorship taking place RIGHT NOW.

Case Citation: Berenson v. Gottlieb, 2026 WL 3135334 (2d Cir. Oct. 20, 2026)

Selected Jawboning Posts

• The Federal Government Used Jawboning to Censor ICE Transparency Initiatives–Rosado v. Bondi
• COVID Jawboning Lawsuit Dismissed (For Now)–Dressen v. Flaherty
• Section 230 (Still) Applies to Contract Breach Claim–NJCCC v. McAleer
• Facebook Defeats Armslist’s Account Termination Lawsuit–Armslist v. Facebook
• Robert F. Kennedy Jr. Is Breaking Internet Law Faster Than I Can Blog It
• Plaintiffs Lack Standing to Sue Over Government Jawboning When Their Evidence is Based on Vibes–Murthy v. Missouri
• Jawboning Defendants Are 6-for-6 in the Ninth Circuit–Hart v. Facebook
• Another Jawboning Case Fails in the 9th Circuit (But a TAFS Judge Doesn’t Like the Biden Administration)–Rogalinksi v. Meta
• Sixth Circuit Dismisses Online Jawboning Case–Changizi v. DHHS
• “Twitter Files” Don’t Help Revive Jawboning Case–Hart v. Facebook
• Another Jawboning Case Fails in the Ninth Circuit–Kennedy v. Warren
• Government Submissions to a Trusted Flagger Program Aren’t Unconstitutional Jawboning–O’Handley v. Weber
• Ninth Circuit Easily Rejects Another Jawboning Case–Huber v. Biden
• Ninth Circuit Easily Rejects Jawboning Claims Against YouTube–Doe v. Google
• Facebook Defeats Jawboning Lawsuit Over COVID Misinformation Removal–Rogalinski v. Meta
• COVID Skeptics’ Anti-Jawboning Lawsuit Fails–Changizi v. Department of HHS
• Facebook & Twitter Defeat Lawsuit Over Account Terminations of COVID/Mask Skeptic–Hart v. Facebook
• Twitter Defeats Trump’s Deplatforming Lawsuit–Trump v. Twitter
• Section 230 Survives Yet Another Constitutional Challenge–Huber v. Biden
• Another Anti-Vaxxer Jawboning Lawsuit Fails–ICAN v. YouTube
• The First Amendment Protects Twitter’s Fact-Checking and Account Suspension Decisions–O’Handley v. Padilla
• One More Time: Facebook Isn’t a State Actor–Atkinson v. Facebook
• Government Jawboning Doesn’t Turn Internet Services into State Actors–Doe v. Google
• Facebook Defeats Lawsuit By Publishers of Vaccine (Mis?)information–Children’s Health Defense v. Facebook
• Another Must-Carry Lawsuit Against YouTube Fails–Daniels v Alphabet
• Congressional Jawboning of Internet Services Isn’t Actionable–AAPS v. Schiff
• Facebook Isn’t a Constructive Public Trust–Cameron Atkinson v. Facebook
• Section 230 Ends Demonetized YouTuber’s Lawsuit–Lewis v. Google

The post Alex Berenson’s Deplatforming Lawsuit Fails in the Second Circuit–Berenson v. Gottlieb appeared first on Technology & Marketing Law Blog.

What do Amy Coney Barrett and Ketanji Brown Jackson do for a living?

The obvious answer is that they are Associate Justices of the United States Supreme Court. That is presumably how they spend most of their working hours. But if we look at their income rather than their job descriptions, the answer becomes less obvious.

In 2025, an Associate Justice earned an annual salary of $303,600. That same year, Justice Jackson received approximately $1.2 million for writing her memoir, while Justice Barrett received nearly $850,000 for hers. Justice Neil Gorsuch earned approximately $300,000 from books, and Justice Sonia Sotomayor, although earning considerably less that year, has collected more than $4 million in book advances and royalties during her time on the Court. By the end of 2025, Jackson’s cumulative book earnings for her memoir have surpassed $4 million. Collectively, the Justices have earned more than $13 million from books since joining the Court.

Apparently, for some Justices, deciding cases is not their most lucrative occupation. Writing books is.

While court ethics rules, including the Supreme Court’s own Code of Conduct, cap outside income at just over $30,000, copyright royalties and book advances are an exception. There is no ceiling whatsoever on how much a Justice can collect from a publisher.

Some might think that it is self-evident that a government official—let alone a Supreme Court Justice—generating most of their income, millions of dollars, from a private conglomerate is obviously a serious problem. But if that doesn’t immediately trouble you, perhaps the extraordinary debacle that unfolded at the Supreme Court this week will.

The Fifth Circuit’s Copyright Adventure

On January 12, 2026, the Fifth Circuit decided Vetter v. Resnik. This trainwreck of an opinion has already been discussed at great length, including on this blog, so I can be brief.

In Vetter, the Fifth Circuit held that when an author terminates a copyright grant under the U.S. Copyright Act, that termination can recapture copyright interests worldwide. To reach that surprising conclusion, the Fifth Circuit effectively treated foreign copyrights as mere extensions of U.S. copyright, rather than as independent rights arising under the copyright laws of other countries.

That was quite a move. Original, certainly. But it also potentially upends decades of understanding about copyright territoriality and the relationship between domestic termination rights and foreign copyright systems.

There is no need to mince words. The decision seems flatly wrong. It has been heavily criticized by many, including at length  by Professor Tyler Ochoa on this very blog (and by me in a forthcoming article).

A Copyright Rescue Mission Goes Shockingly Wrong

Fortunately, the Supreme Court exists to resolve exactly this sort of problem. And in recent years it has shown a certain willingness to rein in the Fifth Circuit’s, shall we say, “creativity.”

There was, however, a practical obstacle. The original losing parties, Robert Resnik and his music publishing business, presumably had little incentive to pursue expensive Supreme Court litigation. Understandably so. The foreign rights associated with one relatively old song might not justify the cost of further litigation, even if the resulting precedent has significant industry-wide implications.

But then the music industry came to the rescue. Or so it seemed.

After the Fifth Circuit decision, the four largest music publishers in the world—Sony, Universal, Warner, and BMG—acquired the disputed copyright interest from Resnik and substituted themselves as defendants. They retained Paul Clement, one of the country’s most prominent Supreme Court advocates, and petitioned the Court to review the decision.

The petition attracted five amicus briefs, including submissions from the Motion Picture Association, the RIAA, international entertainment industry groups, and 20 intellectual property law professors (myself included), led by Professor Ochoa.

The petition was fully briefed at the certiorari stage, including an opposition and reply. This was a serious, coordinated, presumably expensive industry effort to obtain Supreme Court review.

And then came an entirely foreseeable problem. One of the four petitioners was BMG. Unfortunately, BMG belongs to the same corporate family as Penguin Random House, the book publisher of four sitting Supreme Court Justices: Barrett, Gorsuch, Jackson, and Sotomayor.

One hardly needs an elaborate theory of judicial ethics to recognize the potential recusal problem. What makes the industry’s choice particularly baffling is that BMG’s participation was hardly indispensable to assembling a formidable coalition of music publishers. Moreover, this was not an unprecedented problem. In May 2025, the Supreme Court confronted a similar situation in Baker v. Coates, an obscure pro se copyright infringement case involving allegations that Ta-Nehisi Coates and others had copied a writer’s work. There, presumably because Penguin Random House’s parent company was among the defendants, the same four Justices (together with Justice Alito, although probably for different reasons) recused themselves.

But a Supreme Court case cannot handle so many recusals. Under 28 U.S.C. § 1, six Justices are required for a Supreme Court quorum. Thus, in Baker, the Supreme Court could not consider the cert petition on its merits.

The exact same thing—only in a much more meaningful dispute—happened here. The same four Justices who have turned Penguin Random House into an extremely generous source of personal income—Barrett, Gorsuch, Jackson, and Sotomayor—recused themselves. Without a quorum, on October 19, 2026, the Supreme Court disposed of the publishers’ petition in Vetter.

Immediate Implications on Cross-Border Licenses

While this is a rather stunning turn of events, the fallout from this particular episode may prove easier to address than the deeper institutional problem it exposes.

Losing a certiorari petition under such circumstances is an almost laughable blunder. Four massive, sophisticated companies inserted themselves into a dispute, recruited a top Supreme Court litigator, assembled a coalition of amici, and spent considerable resources pursuing review. Yet in doing so, they also created the recusal problem that doomed their efforts. We don’t see this kind of ball-dropping at this level every day.

For now, Vetter remains binding within the Fifth Circuit, but other courts need not follow it. There is already real tension between its reasoning and Second Circuit case law. Yet, the decision potentially casts doubt on the ownership of foreign rights under a large number of existing copyright agreements. And because copyright termination rights become exercisable on a rolling basis, that uncertainty is not confined to any fixed set of agreements. While cross-border copyright termination litigation has historically been uncommon, the resulting uncertainty may still complicate licensing and other transactions and generate further disputes.

It is nevertheless quite possible that the Supreme Court will eventually get another opportunity to clean up this mess. And hopefully, future litigants will verify that nobody on their team belongs to the same corporate empire that has been paying Supreme Court Justices millions of dollars for their literary side gigs.

That is the relatively straightforward part of this story.  But the more troubling implications concern the Supreme Court’s ability to adjudicate copyright disputes generally.

The Supreme Court Has a Publishing Problem

Consider Penguin Random House itself.

This is not some small publisher that happens to have multimillion-dollar contracts with several Supreme Court Justices. It is the world’s largest trade book publishing group, with a vast catalog of copyrighted works. Its parent corporation, Bertelsmann, has substantial interests throughout the media and entertainment industries. Its sister company, BMG, is one of the world’s largest music companies, representing millions of musical compositions.

These businesses are not peripheral participants in the copyright ecosystem. They are among its most important commercial actors.

And, as you might expect, Vetter is not the only copyright litigation involving the Bertelsmann companies. Not even close. And the stakes in some of those other cases are much higher.

Consider just one issue: AI training. In March 2026, BMG sued Anthropic in California, alleging that hundreds of its copyrighted song lyrics had been unlawfully used to train generative AI systems. Penguin Random House authors have also played prominent roles in AI-training litigation. Andrea Bartz, the lead plaintiff in the landmark Bartz v. Anthropic case, is a Penguin Random House author (and the company itself was entitled to claim a share of the gigantic $1.5 billion settlement in this case). So are John Grisham, George R. R. Martin, and Jodi Picoult (and many others)—all named plaintiffs in the pending, massive In re OpenAI litigation in the Southern District of New York.

So, it is not hard to imagine important copyright litigation, possibly dealing with AI training, involving Penguin Random House or BMG reaching the Supreme Court. What happens then? If the four Justices recuse themselves, as they did in Vetter, the Court will again lack a quorum. We would be left with the extraordinary situation where some of the country’s most important copyright disputes cannot receive Supreme Court review because too many Justices have lucrative commercial relationships with a sprawling media group.

That alone is a serious institutional problem.

But the difficulty extends much further than that. What if another AI-training case, even one that does not involve BMG or Penguin Random House or even their authors, reaches the Supreme Court? Presumably, those four Justices would not recuse themselves.

Yet such a case will probably require the Court to address many core questions—such as the transformative nature of AI training, the importance of training licenses, and the validity of the market dilution theory. These questions could have enormous consequences for all AI-training cases and could massively affect Penguin Random House and BMG.

So, we arrive at an uncomfortable situation: the Justices may be unable to hear cases directly involving their publisher but remain free to decide comparable cases that could substantially affect the same publisher’s commercial interests.

Of course, we should not take this argument to an absurd extreme. Judges routinely decide cases that have some indirect effect on industries in which they have some financial interests. But surely there is a difference between having an incidental financial interest in an industry and receiving millions of dollars from a private company whose business may be substantially affected by the legal rules one is asked to announce.

And this brings us to the deeper problem: the legitimacy of the Court itself.

When Copyright’s Judges Are Copyright’s Beneficiaries

The problem is not merely that four Justices have lucrative business relationships with one publishing behemoth. It’s that they’ve become multimillion-dollar beneficiaries of the very copyright system they’re supposed to impartially adjudicate.

As we know, copyright law is not simply a mechanism for ensuring that authors and publishers receive as much money as possible. Instead, as the Supreme Court itself noted, it “reflects a balance of competing claims upon the public interest.”

When Justices receive millions of dollars from exploiting their own copyrighted works, they become financially invested in one side of that broader policy landscape. Even if it doesn’t technically require them to recuse themselves, the appearance problem is glaring. Those Justices collect millions of dollars under contracts with a specific company whose interests frequently come before federal courts.

At some point, the magnitude changes the nature of the relationship. A modest royalty check looks like incidental income from intellectual activity. A multimillion-dollar book deal looks like a major commercial relationship. And it is perfectly reasonable for the public to wonder whether those relationships matter.

Consider, for example, Justice Sonia Sotomayor.

In my view, Sotomayor is an exceptional, important, and brilliant Justice, bringing a distinctive voice and perspective that the Court otherwise might lack. She is also arguably the most experienced of the sitting Justices in copyright law, having practiced intellectual property law before serving on the Southern District of New York and the Second Circuit.

For many years on the Supreme Court, she did not author its major copyright opinions. More recently, however, she wrote the majority opinion in Andy Warhol Foundation v. Goldsmith and a concurrence in Cox Communications v. Sony Music Entertainment. In both, she advanced positions relatively protective of copyright owners. I actually do not necessarily disagree with her. I supported the result in Warhol, although I found its reasoning problematic. And I was very sympathetic to her analysis in Cox.

I’m obviously not suggesting that her book income influenced either opinion. But why should actual corruption be the standard? As noted, since joining the Court, Sotomayor has received more than $4 million in book income from Penguin Random House and its predecessor companies. During that time, she has published multiple books with the publisher. She continues to participate in promotional events for her books, actively supported and publicized by Penguin Random House. This is an ongoing, meaningful business relationship, involving significant labor and significant sums of money, between a prominent Justice and the largest trade book publisher in the U.S. and in the world. During that time, Sotomayor has authored and joined major opinions addressing the scope of copyright protection, including questions with obvious economic implications for copyright owners and their publishers.

Isn’t there something deeply uncomfortable about this arrangement, even if everything is entirely lawful?

This is fundamentally a question of public confidence. Judicial legitimacy requires more than the absence of demonstrable bribery or misconduct. It requires that the public have good reason to believe that cases are being decided independently, without significant financial relationships casting doubt on the process. The appearance of impropriety, after all, is as important as the actual commission of improper acts when it comes to legitimacy.

And to state the obvious: No, none of this is intended to diminish the serious ethics controversies involving Justices Thomas and Alito, including the revelations about undisclosed luxury travel and questions about recusal. Those controversies deserve sustained scrutiny in their own right. Pointing to the problems created by the Justices’ publishing relationships is not an excuse for anyone else’s actions. The Supreme Court seems remarkably capable of having more than one ethics problem at the same time.

In fact, maybe that’s what makes this situation so troubling. Unlike some of the other controversies, the publishing arrangements are largely out in the open. Everything is seemingly above-board.

And yet four Justices have just been unable to participate in a significant copyright case because of their connections to a publishing conglomerate. Meanwhile, those same relationships may raise legitimate concerns about the appearance of impartiality when the Justices do participate in other copyright disputes.

Should Book Income Get a Free Pass?

The Vetter debacle should be an opportunity to reconsider the Justices’ relationships with commercial publishers, the rules governing book income, and the limits of recusal as a solution. In particular, and while I do not pretend to offer a comprehensive framework for reform here, I believe it is quite difficult to continue and defend an ethics regime that limits most forms of outside income while allowing Justices to receive unlimited advances and royalties from commercial publishers.

Treating these payments as merely passive returns on existing intellectual property is something of a fiction. We are talking about enormous sums paid for books that Justices write and promote while on the bench, whose commercial value is tied to the judicial offices they hold. In substance, how is this so different from teaching, which Justices routinely do, but for which their compensation is subject to strict limits? If anything, teaching generally poses fewer concerns about conflicts of interest. And considering that those limits have not prevented Justices from teaching, we should not assume that similar limits would prevent them from writing.

We are entitled to a Court that can resolve important copyright disputes. And we are entitled to a Court whose decisions carry legitimacy, without reasonable doubts arising from its members’ multimillion-dollar relationships with businesses that stand to benefit or lose from those decisions.

The post Four Justices, One Mega-Publisher, and a Supreme Court Copyright Debacle (Guest Blog Post) appeared first on Technology & Marketing Law Blog.

I posted an introduction to the AMA, which I am sharing here:

__

Hi everyone, and thanks for joining this Reddit AMA on web scraping! I thought it would be helpful to start with some introductory remarks.

First, and most importantly, a reminder about the ground rule. Nothing in this AMA constitutes legal advice. I will discuss some general legal principles, but I can’t opine how those principles apply to your circumstances. Further, many of the topics we’re going to discuss are contested, and everyone has different tolerances for legal risk. Before you make any decisions, you need to talk with a lawyer about how the law applies to your circumstance and what level of legal risk you’re willing to accept.

Also, as Ian mentioned, I am focused on US law. The rules outside the US could look quite different!

Some background about me: I am an OG Internet Law expert. I got my first email account in 1991, published my first Internet Law paper in 1993, started practicing Internet Law in 1994, and started teaching an Internet Law course in 1996—a course I still teach today, 30 years later. I have been a full-time law professor since 2002.

As an Internet Law expert who formed his intellectual identity in the 1990s, I still approach Internet Law as a generalist, which was the norm back then. I don’t focus on any one topic of Internet Law. Instead, I still track and pay attention to dozens of Internet Law topics. The scraping/trespass to chattels doctrines are just a subset of my overall portfolio of interests. You can see my academic writings at https://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=332758 Some of my other scholarly priorities include liability for user-generated content (UGC)/Section 230, online trademark law, and emoji law. My forthcoming book project will be about the Internet’s pro-social benefits and how legal regulation jeopardizes those benefits.

Despite my generalist orientation, I find the law of scraping especially interesting. I got my first client question about scraping probably around 1997. When I was in-house, scraping was an evergreen topic. I routinely test on scraping in my Internet Law course. For example, this year’s midterm was based on Vox Marketing Group v. Prodigy Promos, 556 F. Supp. 3d 1280 (2021).

* * *

Part of why I find the topic so fascinating is that I have complicated feelings about scraping from a policy standpoint.

On the one hand, I believe property owners should generally have the ability to decide how their property is used. Scraping can involve unwanted and unconsented uses of a property owner’s property, in ways that can be pernicious.

On the other hand, I believe in access to information. (My mom was a librarian). To the extent that scraping enables better insights, better knowledge, and stronger markets, it’s a social benefit.

Furthermore, the Internet succeeded in part because of its seamless, low-friction, low-transactions-cost navigation. To keep that architecture, browsing the Internet should be a permission-less experience. Anti-scraping doctrines threaten to convert web browsing into a permissioned experience, where there are checkpoints at every site’s perimeter and permission is needed for access. Those barriers threaten to undermine the Internet’s navigability and utility, in ways that will disadvantage all of us.

[Note: Trespass to chattels doctrines are just one way that services can impose permissioned access into web navigation. Copyright law is another. Yet another is the online age authentication mandates sweeping the globe. For more on my opposition to age authentication mandates, see https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5208739 An obvious point: if age authentication checkpoints become ubiquitous in the Internet, that will have major impacts on scraping because most content will live behind age authentication walls].

I have never fully written up my complicated thoughts about these topics, but this piece will give you a sense of my thinking. https://blog.ericgoldman.org/archives/2013/04/rethinking_onli.htm

Here’s how I frame some policy considerations of scraping for students:

On the plus side: Scraping can enable services that create significant social benefits, such as search engines, price comparison bots, and AI modelbuilding. Also, anti-scraping efforts are frequently anti-competitive, i.e., an intent to manufacture or maintain monopolies. In particular, anti-scraping tactics can represent an attempt to obtain IP-like rights, even if the service doesn’t qualify for IP protection. You can see some of these dynamics in the Register.com v. Verio case, where Register.com was using contracts and trespass to chattels to protect non-copyrightable material (WHOIS registrant contact info) to preserve its first-mover-advantage control over that contact information.

On the minus side, scraping can look like a DOS attack (intentionally or not). It can facilitate massive copying, including possibly copyright infringements. And scraping can be a massive invasion of consumer privacy.

So, from a policy standpoint, is scraping a good or bad thing? ¿Por qué no los dos?: Or, to put it into legalese, “it depends.”

* * *

As for the law of scraping, it’s a mess, especially the trespass to chattels doctrines (I summarize them as “TTC”). For about two decades, I taught the Intel v. Hamidi and Register.com v. Verio cases. Around 2024, I felt like those cases no longer reflected the current legal state, so I regretfully took them out of my Internet Law casebook. [For more on the casebook, see https://blog.ericgoldman.org/archives/2026/07/announcing-the-2026-edition-of-my-internet-law-casebook.htm] I have not yet found a good replacement case to add to the casebook, so my casebook coverage of the topic is now adrift. I now teach the X v. Bright Data ruling (the first one granting the motion to dismiss) and it’s not very satisfying. Because there is no good single case or set of cases to teach the topic, and because the law is so unsettled, I tell my students that the topic has become “unteachable.” You can imagine how much they like hearing that LOL.

I teach students to run through a four-step process to determine if/how trespass to chattels applies:

Step 1: does the plaintiff own the affected chattel? Usually, the chattel is the servers or other physical hardware that is used to run the server and connect it to the Internet. The trespass to chattels doctrines are a good illustration of the interplay between the law of physical space and activities taking place in a virtual environment. Students sometimes struggle to wrap their heads around the distinctions between tangible items (i.e., “chattel”) and intangible things, such as data or online accounts, which may be protected by other legal doctrines (copyright or other IP) but don’t legally qualify as the “chattel.”

Step 2: did the defendant “use” the chattel? Usually, sending or receiving packets from a server qualifies.

Created by ChatGPT Dec. 2025

Step 3: was the defendant’s use “unauthorized”? The law might presume that a server owner impliedly permits ordinary Internet traffic merely by making the choice to connect to the open Internet. How can the server owners remove this implied permission? This is one of the areas that is unteachable. There is no clear answer to how a server owner can structure this dialogue with Internet users. Some of the possible tools that a server owner MIGHT be able to use to remove any implied authorization to use their chattel:

• TOSes
• C&D letters
• Technical access controls like registration walls
• Technical access controls like IP address blocking
• Signals such as robot exclusion headers

But, none of these mechanisms are guaranteed to work. This is an area of doctrinal chaos. I’m sure we’ll spend some time discussing if and when those efforts are effective. In general, for each bullet, I have more questions than answers.

Step 4: did the defendant’s unauthorized use create legally recognizable harm? This is also an unteachable aspect because, again, I typically have more questions than answers about what harm will suffice. The court’s flip-flop in the X v. Bright Data case is a good example of how unpredictable this topic is.

* * *

For my students, I summarize the law of scraping with the following tagline: “Scraping is ubiquitous but not clearly legal.” This breaks my students’ brains. If it’s so ubiquitous, it must be clearly legal…right?

In order for a scraping campaign to be legal, we have to navigate ALL of the following doctrines:

• Trespass to chattels. I tell my students that they should always discuss all three of these doctrines as a troika:
  • Federal statutes, especially the Computer Fraud and Abuse Act (CFAA).
  • State statutes, such as the state computer crime laws. Here in California, we focus on Cal. Penal Code 502.
  • Common law trespass to chattels, such as the Hamidi and Bright Data cases.
• Copyright law (106). I spend some time explaining why ordinary web browsing might be copyright infringement because copyrighted material necessarily is downloaded into the browser’s RAM. The students really don’t like that discussion LOL. If web browsing is infringement, then scraping would be as well. In addition to the initial downloading, scrapers might encounter further 106 trouble with republishing some or all of the scraped material.
• Copyright law DMCA (1201/1202). I don’t cover this angle in class yet because it’s too underdeveloped and chaotic, but I’m sure we’ll spend some time on 1201 today. To give you a preview, the 1201 rulings to date break my brain, and I have more questions than answers.
• Many TOSes contain anti-scraping provisions, but it’s not clear if the TOSes are properly formed. In my Internet Law course, we spend a full 2 weeks discussing TOS formation. The law evolved a lot in the 2025 decisions in Chabolla and Godun in the 9th Circuit.

There may be other legal doctrines raised by scraping, such as trademarks and privacy rights.

So here’s why I say that scraping is not clearly legal. In order for a scraping campaign to be permissible, it must survive each and every one of the doctrines enumerated above. If the campaign fails any one of the doctrines, it’s not legal and the scraper can face significant consequences (including, in some cases, criminal liability). It is daunting to build a scraping campaign that can successfully navigate so many legal doctrines perfectly.

* * *

My apologies for such a long-winded introduction (although, I remind you that I’m a law professor, so loquaciousness isn’t really that surprising, is it?) The law of scraping is so complicated and multi-faceted that it’s easy for the discussion to go down rabbitholes or for conversants to talk past each other. I hope this intro provides some helpful structure and context for my more detailed comments.

Thanks again for engaging with the AMA. A blanket apology in advance for all of the times I answer “it depends,” “I don’t know,” or “that’s a good question.” And in all cases, don’t take my word as the final one—talk with your lawyer if you are making any decisions!

The post My Reddit AMA on the Law of Web Scraping appeared first on Technology & Marketing Law Blog.

The opinion summarizes:

As part of its ubiquitous ride-hailing service, Uber permits a user to order “guest rides” for third parties remotely. Guest riders can travel unaccompanied by the user and need never have downloaded Uber’s app orcreated an account themselves. This case concerns what legal obligations a guest rider owes to Uber by virtue of using the service.

In this case, Cheryl ordered an Uber ride for her husband Carroll. Carroll suffered significant personal injuries on the ride. Uber sought to direct Carroll’s claims to arbitration. The court says Carroll never agreed to Uber’s TOS, so no arbitration.

The court summarizes its legal conclusion:

Uber relies on its having sent Carroll a text message before the guest ride that included a hyperlink to the Terms and stated that taking the ride would manifest his assent to them. But Uber cannot show that Carroll ever saw its unsolicited message, and contract law imposes no duty to read a contract that one does not know exists. Uber alternatively submits that Carroll could be separately bound to arbitrate his claims under Cheryl’s contract with Uber. Carroll does not seek to enforce his wife’s contract, however, and he had no more notice of its terms when he took the guest ride.

Why Didn’t the Text Message Work?

Uber texted Carroll:

Your Uber from Cheryl is arriving in 10 minutes. \n \nBy taking this trip, you agree to the Uber Terms of Use & Privacy Policy: t.uber.com/lgl. \n n\Track here/nhttps://trip.uber.com/Wu8vBbIYEf3A \n\nReply STOP to opt out from SMS notifications.

The court says that Carroll did not have inquiry notice:

Under the design of Uber’s guest-rider system, it never required Carroll to click on an “I accept” button or its equivalent, or to otherwise expressly confirm his acceptance of the Terms. Nor was Carroll presented with a website- or app-based interface that contained a link to the Terms and stated that taking the ride would manifest his assent to them. Instead, Uber simply sent Carroll a text message with that information and anticipated he would see it before taking the ride. But individuals are not on inquiry notice when they must “ferret out hyperlinks”—even when they are already using the apps or webpages on which the links appear….In focusing on the layout and language of the text message it sent Carroll, Uber skips over a critical difference between this case and the mine run of internet-contracting cases: Carroll never saw the message, so it is irrelevant whether he would have been on inquiry notice from the message had he seen it.

Uber suggested that Carroll surely saw its text. The court responds:

The system Uber designed did not require guest riders to read the text messages Uber sent them before using the service. Nor did its system as a practical matter require guest riders to open the messages to obtain information about their rides…

we disagree that an objectively reasonable person in Carroll’s position necessarily would have read the text message from Uber before taking the guest ride….At least until they enter the vehicle, the guest rider’s relationship to Uber’s service is entirely passive; they need never have interacted with the company in any way. In that context, it is far from apparent that a reasonable person would expect to receive an unsolicited text message calling for them to examine voluminous contract terms in the short window before their ride arrives. And although Carroll’s practice of never reading text messages may be idiosyncratic, there is no legal duty to perpetually monitor one’s phone for incoming messages that may contain contract terms interspersed with other text.

“Never reading text messages” sounds like aspirational goals for many of us.

Other Ways of Binding Carroll

Carroll isn’t a third-party beneficiary of Cheryl’s TOS with Uber. “Carroll is not seeking to enforce Cheryl’s contract with Uber, nor did he otherwise consent to be bound by its terms….if Carroll was an intended third-party beneficiary of the contract between Cheryl and Uber, that would simply mean he could sue Uber to enforce its obligations to Cheryl. That is a far cry from saying he incurred his own obligations to Uber from a contract to which he was not a party.”

Also, “Even assuming the guest ride counts as a direct benefit that Carroll accepted under Cheryl’s contract with Uber, Uber’s [equitable estoppel] argument still fails. We have explained that Carroll lacked actual or constructive notice of the terms on which that benefit was being offered, so he did not “knowingly” accept any benefit.”

What Could Uber Do Differently?

In my Internet Law course, I describe a phenomenon of TOS formation “leaks,” where people can reach their desired goal without going through the otherwise mandatory TOS formation process. Uber guests are an example of such a leak in Uber’s TOS formation process.

Uber will always have a problem binding every rider to a TOS (and thus arbitration). If an Uber member requests a ride that includes unrelated individuals, Uber doesn’t know who the additional passengers are or have any direct contact with them sufficient to form a TOS with them. Thus, Uber inevitably has limited arguments to divert non-members into arbitration.

In this respect, Uber is in a better position to bind Uber “Guests.” If Uber gets an order for a guest, it could cancel the ride unless the guest agrees to the TOS, such as requring the guest to reply to its text (as the court says, “Uber, for instance, could require guest riders to confirm receipt of the Terms by responding to its text message”) or clicking through some TOS offer screen before the guest order is confirmed. However, I doubt these options would be very popular. For example, guests may not have a cellphone or may have some infirmity that explains why third parties are handling their affairs.

Alternatively, Uber could put the burden on drivers to ensure that everyone who gets in the car has agreed to Uber’s terms somewhere/somehow.

First, Uber could require everyone ordering a ride to specify the identity of riders and their contact info, and then confirm them like the guest protocol I mentioned above. This would require drivers to then block access to guests who didn’t assent to the TOS. I don’t think anyone would love this solution, including drivers who might face physical threats for denying rides or potentially leave people stranded in dangerous situations.

Second, Uber could put some kind of TOS disclosure on car windows, such as “by entering this vehicle, you agree to the TOS.” (Call it a…window-wrap?). Putting aside the legal dubiousness of a window-wrap, it would require drivers to implement this properly on their cars, and the odds of that are prohibitve. (Also, a car usually has multiple doors, and each one would need the disclosure if the window-wrap were to work against everyone).

In sum, it’s quite possible that Uber may never be able to bind all riders to arbitration. Then again, I don’t recall taxi services in their day routinely being able to bind their riders to arbitration either.

Case Citation: Walker v. Uber Technologies, Inc., No. 24-7154 (D.C. Cir. Aug. 28, 2026)

* * *

Schlueter-Beckner v. SimpliSafe, Inc., No. 25-4933 (9th Cir. Aug. 31, 2026)

SimpliSafe’s TOS formation via a sign-in-wrap failed for lack of visibility. The Ninth Circuit doesn’t show the screen, but the lower court opinion included this:

The Ninth Circuit first explains why it focuses only on the bottom “place order” widget, not the one on the right:

Because the fill-in items that required action and attention from the consumer were all in the left-hand column, the overall design of the page deemphasized the right column and drew focus toward the “Place order” button at the bottom of the left column. This emphasis is reinforced by the fact that the preceding “Shipping” webpage in the purchase process similarly placed the fill-in items on the left-hand side, with a similar blue action button at the bottom. Moreover, SimpliSafe presented no evidence to show that Plaintiffs must have used the “Place order” button on the right-hand side or must have read the disclosure on that side to finalize their purchase. Therefore, any insufficiency with the bottom disclosure—which may have been the only disclosure the user saw—would mean SimpliSafe failed to show its webpage provided reasonably conspicuous notice.

Then, the Ninth Circuit says that the bottom TOS offer isn’t prominent enough:

The disclosure’s placement, font size, and font color drew the user’s attention away from the disclosure. There was significant white space between the disclosure and the “Place order” button; the text of the “Place order” button was seemingly twice the size of the disclosure’s text; and the “Place order” button contrasted more against the white background than the disclosure’s gray text….

Go back and double-check the white space between the TOS offer and acceptance button. Seriously?

The court continues:

Above the bottom disclosure, the purchasing webpage includes a line of text warning users their credit card information would be automatically saved to the account for future use. This warning is in larger font than the disclosure, italicized, and in navy text against the light gray background. As a result of these design choices, the credit card notice contrasts starkly on the page and draws the eye. So, while SimpliSafe utilized design elements to ensure the user saw the warning aboutretaining credit card information, “they have not done the same to alert a consumerto the Terms of Use.”

Would you say that the credit card warning was so much more prominent than the TOS offer?

Some plaintiffs also signed up for alarm monitoring, and that signup process failed too. The TOS offer labeled the TOS “SimpliSafe Terms and Consent to Communicate, for Notifications and Alerts.” The court said that sounds like the TOS only related to SimpliSafe’s messages, not a more comprehensive set of TOS terms.

Also, the TOS offer and acceptance button didn’t match:

the disclosure stated “[b]y submitting this number” Plaintiffs agreed to the terms, but the action button on the webform only stated “Next.” This mismatch of terms could easily confuse users.

As I”ve said before:

* * *

Uhl v. Roblox Corp., No. 25-5057 (9th Cir. Sept. 3, 2026)

Roblox’s TOS contains an arbitration clause but “Roblox did not move to compel arbitration—for months on end.” Roblox claimed it needed to know the plaintiff’s username to determine if the case was arbitrable, but Roblox didn’t seek any discovery to resolve this ambiguity. Instead, it litigated the preliminary stages of the case in court, including a removal motion, a remand, and a substantive motion to dismiss.

This makes it look like Roblox sandbagged the arbitration issue. The majority summarizes:

a party cannot ask the district court to dismiss a complaint on the merits, holding in reserve a claimed right to arbitrate that will be exercised only if it does not like the district court’s decision.

Judge Bumatay (a TAFS judge) dissents, saying the law wasn’t settled or clear that Roblox was waiving its arbitration rights by proceeding as it did.

The post Uber Guests Aren’t Bound By Uber’s TOS–Walker v. Uber appeared first on Technology & Marketing Law Blog.

In response, the firm sued her for defamation and breach of contract. The firm enumerated seven allegedly defamatory statements, including “They are either scam artists or complete crooks” and “They are not tough, they are not effective and they will rob you blind.” The lower court held these statements were nonactionable hyperbole and the author’s opinion. Other claims failed because the firm didn’t show they were false.

With respect to the contract breach, Section 16 of the firm’s fee agreement said the parties “agree not to directly or indirectly disparage each other either electronically or otherwise.” Because of this, Yelp has flagged the firm’s Yelp business page with a “Questionable Legal Threats” warning:

The link goes to the full text of the Superior Court opinion in this case.

As every lawyer knows, suing a client entails substantial risk of blowback, including possibly a malpractice claim and a complaint to the state bar. Here, the law firm gets other unwanted consequences. Sedgwick successfully brought an anti-SLAPP motion to strike, which means the firm will be paying her legal fees.

The appeals court upholds the anti-SLAPP dismissal, saying “we readily agree with the trial court that the Yelp review criticizing the firm’s representation of Sedgwick is protected activity.” The appeals court adds that “calling someone crooked is ‘merely rhetorical and hyperbolic language’ and not defamatory.”

Of particular note is the court’s treatment of the fee agreement’s non-disparagement clause. The law firm equated disparagement with defamation, so the court said the contract breach claim failed when there was no defamation. The appeals court adds:

Section 16 violates Civil Code section 1670.8. That statute provides: “A contract … for the sale … of consumer goods or services may not include a provision waiving the consumer’s right to make any statement regarding the seller … or its employees or agents, or concerning the goods or services.” By its plain language, it precludes a provision that, like Section 16, purports to require Sedgwick “ ‘not to directly or indirectly disparage’ ” the firm.

We do not share the firm’s concern that this statute “would negate existing law and permit consumers to defame businesses … with impunity.” The statute cannot be reasonably read to immunize defamation. In our context, the statute does not prevent the firm from suing for defamation, if the firm could prove defamation sufficiently to survive an anti-SLAPP motion. The statute merely prevents the firm from including Section 16 in its fee agreement—or thereafter enforcing it.

I haven’t exhaustively researched the issue, but I couldn’t think of a prior court finding a 1670.8 violation. It’s 2026, and I can’t believe anyone–especially a law firm–still doesn’t know that contractual efforts to restrict consumer reviews are no bueno.

I’ve complained about the spate of bogus 1670.8 “drive-by” litigation that is clogging our courts, but here is a real-life bona fide example of an actual genuine 1670.8 violation. Review suppression like this is what the drive-by lawyers should be enforcing, not their current tendentious and tortured reading of TOSes.

The argument that the law firm needs a contractual anti-disparagement clause to prevent defamation is specious. If a client posts a defamatory review, defamation law applies–no contract overlay needed. I discuss that issue more in this post.

It would have been more interesting if the firm had claimed that the contract anti-disparagement clause applied even if there wasn’t any defamation. But 1670.8 would kick in, mooting that claim as unenforceable.

Case Citation: LS Carlson Law PC v. Sedgwick, 2026 WL 2410113 (Cal. App. Ct. August 17, 2026)

BONUS: Greenwich Building Company v. Hoots, 2026 WL 2408839 (N.Y. Supreme Ct. August 11, 2026). A real estate company sued a property buyer for negative reviews on Google and Houzz. The court finds the lawsuit was covered by New York’s anti-SLAPP laws. The court says “online review forums such as Google and Houzz have consistently been held to be public forums under New York’s anti-SLAPP law…The statements are of interest to members of the community who may consider buying or renting an apartment at 120 Java Street.”

The court then says the plaintiff showed a substantial basis for its defamation claim because there is some confusion about the identity of parties (the buyer’s complaints may relate to a different, though possibly, related company).

And yet, the plaintiff didn’t adequately show the buyer had actual malice: “Defendant left his reviews on plaintiff’s Google and Houzz pages after diligent pedestrian online research….defendant’s statements specifically on those online review pages appear to have been the result of an honest mistake and that defendant left his reviews on plaintiff’s pages after a good faith attempt to find the correct pages to leave such reviews….This court too chooses to err on the side of non-actionability in this case so as to promote the public’s right to free speech.”

As a result, the case is dismissed, and the defense will get its attorneys’ fees.

Selected Blog Posts Regarding 1670.8

• Another Court Shuts the Door on “Yelp Law” Drive-By Lawsuits–Arterberry v. Peet’s
• Appeals Court Shuts Down Bogus Private Enforcements of the ‘Yelp Law’–Moss v. GoDaddy
• Catching Up on the Bogus “Yelp Law” Litigation Campaign–Tao v. Uniqlo
• Amazon Must Defend “Yelp Law” Claim–Ramos v. Amazon
• Courts Are Rejecting Attempts to Weaponize Laws That Protect Consumer Reviews
• Businesses Cannot Contractually Ban “Abusive” Consumer Reviews
• California Tells Businesses: Stop Trying To Ban Consumer Reviews (Forbes Cross-Post)

The post Lawyer’s Fee Agreement Gagging Client Reviews Is Illegal–LS Carlson Law v. Sedgwick appeared first on Technology & Marketing Law Blog.

“Grok, is it bad when TOS formation fails?”

“Grok, what are best practices for TOS formation and does our current TOS formation process satisfy those?”

* * *

The plaintiff claims Grok “disclosed private and confidential information.” Grok sought to transfer the case to N.D. Tex. per its TOS forum selection clause. Everyone agrees the case should transfer if the TOS was properly formed. It was not.

The Wrap Taxonomy

Here is the applicable TOS formation screen at account signup. Everyone agrees this is a sign-in-wrap.

Here is the screenshot when Grok users submit queries:

The court says the chat screen is closest to a sign-in-wrap as well.

Thus, the court follows the now-standard Chabolla framework for reviewing sign-in-wraps.

Transaction Context

Unhelpfully, the parties obviously didn’t think this factor was important:

the Parties devote little space to addressing the transactional context at play here. Defendant offers one sentence in a footnote in its reply brief, citing to Oberstein and contending that “signing up for an account is the type of transaction that contemplates an ongoing relationship.” Plaintiff does not discuss this issue.

This raises another question for Grok: “Grok: is it good when a litigant fails to provide advocacy for one of the elements the court will analyze?”

The court says the plaintiff should have assumed there were applicable terms:

The first factor is that, although an account is “not required” to use Grok, Plaintiff chose to make one anyway. The second is that, after making that account, Plaintiff used Grok repeatedly “throughout 2025 and 2026” to “enter[] queries related to sensitive information about finances, investment strategy, private health conditions, business projects, and other private information.” In the Court’s view, a user who affirmatively chooses to make an account when doing so is optional and recurrently uses the corresponding web service should reasonably contemplate “some sort of continuing relationship” accompanied by terms and conditions. That Plaintiff used Grok to discuss sensitive personal information also weighs in favor of the notice requirement.

From my perspective, anyone who voluntarily chooses to use Grok extensively has placed themselves into a caveat emptor situation. It’s like the Farmer and the Snake parable.

Reasonable Notice on the Sign-Up Screen

The court then turns to the visibility of the TOS offer language, starting first with the sign-up screen. The court says there is a visual discontinuity between the TOS offer and the acceptance buttons:

although a reasonable user is presented with four sign-up methods, he ultimately selects only one. Having made that selection, the other sign-up options necessarily become “superfluous.” Even if a user chose the sign-up option closest to the bottom of the webpage (the option to “Sign up with Google”), Defendant’s contention that the notice is “positioned immediately below” is still incorrect. The language that is directly below the “Sign up with Google” button reads: “Already have an account? Sign in[.]” A reasonable user who is on Defendant’s sign-up page to sign up for a Grok account, which is how Defendant contends Plaintiff agreed to its terms, would likely see that text and read no further because, by definition, it does not pertain to him

The court is doing some tricky but important work here. Basically, she’s saying that once a user encounters language that isn’t applicable to them, they will stop reading and proceed. So here the “Already have an account? Sign in” language acts as a visual barrier that moots everything below it for a user who is signing up for the first time. Perhaps if the TOS offer language were closer to the buttons and more prominent, it could have overcome that visual barrier. But putting the language above the sign-up buttons would have avoided this limitation. (Or, better yet, two clicks).

The court doesn’t like the TOS offer language presentation either:

The contrast between the white hyperlink and the black background is notable, but it says nothing of the contrast between the hyperlink and the text immediately adjacent to it….the difference between light gray and white is a matter of degree and hardly one that maximizes visibility. Moreover, any distinction that this color variation provides is offset by the fact that other non-clickable text such as the “Create your account” heading is also in white…. the single colorful feature—the multi-colored Google logo on the “Sign up with Google” button—is near the center of the screen, likely drawing a reasonable user’s attention more than the monochromatic notice text at the bottom of the page does

I guess the judge doesn’t love monochromatic screens as much as Musk does?

The court summarizes:

Although the transactional context suggests that Plaintiff should have expected to be bound by contractual terms, the Court finds that the sign-up screen notice is not reasonably conspicuous because of “the notice’s distance from relevant action items, its placement outside of the user’s natural flow, and its font—notably timid in both size and color.”

I’m still waiting for courts to clarify what the “user’s natural flow” means. It brings to mind the great Dave Mason song (RIP).

Chat Screen TOS Formation

The chat screen formation evaluation doesn’t fare any better (and maybe a little worse?):

it is patently untrue that the notice is “directly beneath the query box.” Nothing is directly beneath the query box except a lot of empty space. Moreover, the query box is the only action item on the chat page, and as such a user’s “most obvious and natural next step” is to input text into that box. Because the query box is on the top third of the page and no additional graphics or text give the user any reason to look beyond it, the notice at the bottom of the website is “buried” beyond “the periphery” of where a user would be looking.

A “lot of empty space” “buries” the TOS offer language.

The court also noticed the lack of underlining to signal the link: “The only factor distinguishing the hyperlinked text and the adjacent text is the font color, and, as the Court noted, white and light gray are not “contrasting font colors.””

X pointed to a long list of court cases upholding its TOS formation. The court says there were factual differences in all of those cases. I imagine this will be contested on appeal.

Implications

One more question for Grok: “Grok, after reading this opinion, please suggest the six most important changes we should make to increase the likelihood of TOS formation.” The list would be trivially easy to implement; and there’s no reason Grok waited this long to figure it out.

This opinion is a good example of how Chabolla expedites the TOS formation failure in screens like this. This implementation was possibly marginally passable pre-Chabolla. Now, in my opinion, it’s clearly deficient, but in easily avoidable ways.

Case Citation: Skaggs v. X.ai, LLC, 2026 WL 2352969 (N.D. Cal. August 13, 2026). Defense counsel in this case is Kenneth Michael Trujillo-Jamison, Willenken LLP.

* * *

BONUS: Trimble v Entrata, Inc., No. 25-1975 (4th Cir. Aug 11, 2026).

This case involves the RentalPortal service, which allows tenants to pay their rent but charges a convenience fee for the pleasure of doing so. It sounds like RentalPortal used a clickwrap: “Although users could not finalize their rent payments until they had checked the box indicating their agreement to the Terms, they were not required to click on, scroll through, or otherwise actually review the hyperlinked Terms.”

Tenants sued for the convenience fee collection. RentalPortal moved to send the case to arbitration. TOS formation isn’t at issue. Instead, the plaintiffs challenge consideration.

In an echo from Harris v. Blockbuster from 2009 and the more recent Johnson case, a majority of the panel (over a dissent by Judge Rushing) says the arbitration clause is illusory because “Appellant’s Terms include a Change Clause that expressly gives it the unfettered discretion to unilaterally modify the arbitration agreement without any meaningful limitation.”

The change clause reads:

At [ResidentPortal], [users] are bound by the version of this Agreement that is in effect on the date of [the user’s] visit. This Agreement may change from time to time, so please review it when you visit [ResidentPortal].

The majority says:

Under a plain reading, a person who visits ResidentPortal is automatically bound by the Change Clause the moment ResidentPortal is accessed. The fact that users must click the assent button as part of the user interface in order to make rental payments does not alter the fact that any user paying rent is already bound by any modified terms merely by visiting ResidentPortal in the first instance. We have repeatedly held this type of “change in terms” clause to constitute an illusory promise under Maryland law….

there is no language in the Change Clause here requiring Appellant to provide its users with notice before or even after modifying the terms….

the Notices Clause states that any notices Appellant chooses to provide will be posted to ResidentPortal. But it also provides that the Terms automatically apply if and when Appellee visits ResidentPortal. So, there was no way for Appellee to receive notice of modifications or review them without automatically assenting to them. Therefore, the purported ability of Appellee to opt out of accessing the website and thereby decline the modified terms is illusory because it exists solely at Appellant’s pleasure…

it is entirely in Appellant’s discretion to modify the Terms, and the only restriction on that authority is Appellee’s choice to access ResidentPortal to pay her rent — which is no choice at all, as she is automatically bound by the version of the Terms posted the moment she accesses the website, even though at the time she accesses it, she would not know whether there is any posted notice of new Terms.

The majority concludes:

because the Terms here “allow[ed] [Appellant] to escape all of its contractual obligations at will” and did not provide Appellee with effective notice of any modifications, the agreement to arbitrate here is precisely the type of promise we have previously found to be illusory under Maryland law

Arbitration denied.

Given what we learned in Harris v. Blockbuster 17 years ago, it’s incredible that in 2026 we’re seeing these kinds of TOS amendment implementations. Any TOS amendment clause that purports to permit TOS amendments without actual notice to the users is a toxic provision for the entire TOS. Here, RentalPortal can easily form a new contract through a clickwrap each time the user pays rent, so this outcome was incredibly easy to avoid.

The post Grok’s TOS Formation Failed–Skaggs v. X.ai appeared first on Technology & Marketing Law Blog.

Consumer Consent

LinkedIn’s own disclosures may be sufficient (see LB v. LinkedIn in my prior post), but its partners’ disclosures are not. Covered California’s disclosures expressly say “no personal information is sold or distributed,” and Nevada Health Link’s disclosures do not mention sharing data with third parties.

Furthermore, “Courts have often found a lack of consent when the type of data collected, though conceivably included in the disclosure, falls outside the plaintiffs’ reasonable expectations.” LinkedIn capturing consumers’  “medication and dosage” would be outside such expectaions. The complaint against Covered California was not specific enough about its data captures, but that can be fixed in an amended complaint.

LinkedIn’s Intent to Intercept

Johnson alleges that LinkedIn intentionally intercepted her and class members’ communications regarding their medical care through the Insight Tag, which Johnson alleges LinkedIn designed to collect all interactions with the Nevada Health Link website. Johnson also alleges that the issue of health information being shared via tracking technology is so well known in the technology industry that LinkedIn should have known its Insight Tag was likely collecting personal health information. Finally, Johnson alleges that LinkedIn monetized the intercepted personal health information in the form of targeted advertising, which is most of LinkedIn’s revenue. Though the inference of intent may be relatively weak in light of LinkedIn’s policies to the contrary and the absence of specific facts concerning how LinkedIn uses the data it allegedly receives from the Insight Tag, the Court finds these allegations sufficient to survive a motion to dismiss

So…the court says LinkedIn has sufficient “intent” to intercept sensitive consumer data even if its contract with customers restricted such efforts, because LinkedIn should have known data leakages would occur? That doesn’t sound like “intent” to me. See the Pharmatrak case from 2003…

LinkedIn argued that the Nevada Health Link is a government agency, so it should be entitled to the presumption of regularity, i.e., that it followed its own policies and the LinkedIn agreement. The court says that presumption, if it applies at all, is more of an evidentiary burden, not a legal matter that supports motions to dismiss. Plus, Johnson alleged a wrongful data collection of their data, regardless of the entities’ general policies.

Intrusion Upon Seclusion

LinkedIn fails to provide any support for its argument that the “amount” of information disclosed, the public availability of the website, or the optional nature of some forms are relevant to the degree of intrusion or otherwise cause the context, circumstances, and setting factors to tip in LinkedIn’s favor. For purposes of the motion to dismiss, the Court finds Johnson’s allegations that LinkedIn’s unauthorized interception of information regarding her medication and dosage was “highly offensive” conduct sufficient.

Based on my unscientific review of the Pixel cases, it seems like the lawsuits against healthcare entities frequently survive motions to dismiss. It will be an entirely different matter for plaintiffs to win the cases and defend those rulings on appeal.

Case Citation: Doe v. LinkedIn Corp., 2026 WL 2199526 (N.D. Cal. July 30, 2026)

More Posts on the Pixel Cases and Related Claims

• Third Circuit Rejects a Meta Pixels Case–Cole v. Quest Diagnostics
• The Wrap Taxonomy Vexes the Judge in the LinkedIn Insight Tag Cases
• The Second Circuit Is Done With Meta Pixels VPPA Cases–Hughes v. NFL
• Court Rejects Trespass to Chattels Claim Over Placing Cookies–Doe v. Tenet
• Clickwrap Formed Even When a Consumer Has Limited Time to Act–Washington v. Flixbus
• Meta Pixels Case Dismissed by Second Circuit–Solomon v. Triller
• Leaky TOS Formation = No TOS Formation–Snyder v. G6
• The VPPA May Be a Dinosaur Statute, But It’s Very Much Alive in the Second Circuit–Salazar v. NBA
• Pixel Case Against Google “Jumps the Shark”–Doe I v. Google (Catch Up Post)
• Think You Understand Online Trespass to Chattels Law? Think Again–In re Meta Healthcare Pixels
• More Chaos in the Law of Online Contract Formation
• Privacy Lawsuit Based on Website Tracking by Service Provider Trimmed

The post LinkedIn’s Insight Tag Creates Legal Entanglements–Doe v. LinkedIn appeared first on Technology & Marketing Law Blog.

Created by ChatGPT Dec. 2025

by guest blogger Kieran McCarthy

I recently described the district court’s decision in the case of Amazon v. Perplexity as “a shockingly poor effort to grapple with CFAA applicability to agentic AI technology.”

The Ninth Circuit appears to have agreed.

This week, the Ninth Circuit panel published its opinion vacating the district court’s preliminary injunction and remanding for further proceedings. Amazon.com Services, LLC v. Perplexity AI, Inc., 2026 WL 2237587 (9th Cir. Aug. 4, 2026).

The panel determined that Amazon was not entitled to a preliminary injunction because it was not likely to succeed on the merits of its CFAA claim. The court concluded that Perplexity did not “access” Amazon’s computers within the meaning of the CFAA, but rather it was the user who accessed Amazon’s computers. The panel also concluded that Amazon was equally unlikely to succeed on the merits of its CDAFA claim for the same reasons.

The panel also concluded that the remaining equitable factors favored Perplexity.

The Court kept its opinion narrow and specifically disclaimed the applicability of this opinion to other facts or other laws related to data access and agentic AI. But this was a nonetheless important step in distinguishing the infamous Power Ventures opinion, at least in the context of agentic AI.

—

I described the facts in detail in my prior write up of the district court’s decision, so I won’t rehash them here.

But unlike at the district court, the Ninth Circuit panel did not ignore the key factual pattern that makes this case so interesting. What’s unique and interesting about this case is that it involves a fundamentally new technology, in the form of Agentic AI, acting on the instructions of end users.

The panel wrote:

Agentic AI is an emerging technology. There is thus little to no existing caselaw directly dealing with how to ascribe responsibility for AI agents like the Assistant, let alone caselaw specifically dealing with agentic AI in the CFAA context. And CFAA cases dealing with more established technologies do not provide a perfect analogue to the case at hand.

Id. at 12.

The panel then went on to distinguish the facts of both Power Ventures and BrandTotal (both cases involving Facebook, ironically enough).

And the key distinguishing fact, according to the panel, was that on the facts presented to the Court, it was not Perplexity that was accessing Amazon’s computers, but rather the end users through their computers.

Under the CFAA’s plain language, Perplexity’s Comet Assistant itself cannot “access” Amazon’s servers.

According to the Court:

However advanced the Assistant currently is, it is a tool, not a person for statutory purposes. See 18 U.S.C. § 921(a)(1) (“The term . . . ‘whoever’ include[s] any individual, corporation, company, association, firm, partnership, society, or joint stock company.”); see also Whoever, Cambridge English Dictionary, [https://perma.cc/YY3TVTJF] (last visited July 16, 2026) (“[T]he person who” (emphasis added)).

Id. at 15.

Since it is the user using the Assistant tool, Perplexity’s not accessing Amazon’s servers in the first place. And the entire CFAA case unwinds after it reaches that conclusion.

The Court further reinforced its conclusion with the rule of lenity (which courts that enforce the CFAA always seem to forget). The Court was concerned that under Amazon’s theory of liability, end users could be exposed to criminal liability for using agents. Given the rapid adoption of this new technology, the Court was concerned about that approach.

But with just as much caution, the Court was equally quick to reiterate what the opinion was not.

We do not establish a new legal regime governing agentic AI. We do not address whether in other contexts, including tort claims, Perplexity can avoid liability for the Assistant’s actions. Our holding here is limited to “access” as contemplated by the CFAA and as applied to the Assistant’s interactions with Amazon.com on the record before us, not the broader landscape surrounding agentic AI.

Id. at 17.

—

The Court’s reasoning was largely identical in the context of the CDAFA, even though, as I noted in the prior district court opinion write up, the CDAFA is broader than the CFAA. Even though the CDAFA is broader, the “focus of the inquiry is still on the person accessing or causing the access.” Id. at 18. And since the panel concluded that it was not Perplexity that accessed Amazon’s servers, but rather end users through their own computers, that was that.

—

Once the panel determined that Amazon was unlikely to succeed on the merits, the district court’s thin and largely circular conclusions on the other equitable factors unraveled quickly.

In the end, the Ninth Circuit panel kept its opinion tight and focused on the primary issue in the case: whether Perplexity’s agentic AI tool was a “person” who accessed a site within the meaning of the CFAA. It is not, and the case has been remanded for further proceedings in light of that conclusion.

The case avoiding sweeping pronouncements and it was not the straight-up revocation of Power Ventures that some might have hoped for. Given the other legal issues presented, Perplexity might not be out of the woods just yet, but it’s fair to say that the Ninth Circuit vacated bad legal precedent and replaced it with a better one. And for that we can be thankful.

* * *

Eric’s Comments

In the copyright module of my Internet Law course, I ask students a simple question: is web browsing infringement and, if so, by whom? I use the Cablevision case to review the question. (In that case, the court held that for a DVR-as-a-service business, the users pressed the button to make the storage copies). The whodunit question blows students’ minds because it’s 2026 and they assume these issues have long been resolved. As this case shows, courts are still unsure who takes the legally significant actions online.

The ruling also brought to mind the Ticketmaster v. RMG case (the 2007 Hannah Montana case), which I still teach today. The facts in that case remain murky to this day, but I always interpreted RMG as handing off its software to its ticket broker customers, who then operated the software to snipe scarce event tickets, but with some ongoing facilitation from RMG (like relaying ticket purchase requests through a phalanx of RMG-associated IP addresses to reduce blockage). Ticketmaster successfully prevailed over RMG in that case on several legal theories, but does this court’s reasoning suggest that Ticketmaster’s redress should come solely from RMG’s customers, not RMG? Or does the RMG case suggest Amazon ought to win against Perplexity, just under other legal theories?

In support of that latter prospect, the opinion invites courts to more closely scrutinize exactly what the providers of agentic AI software do to facilitate their users’ activities. If they cross some undefined and currently invisible line, the software provider will transform into the legally relevant actor. That might harmonize this ruling with the RMG and Cablevision cases, but at the cost of inconsistent judicial outcomes and high defense costs.

As with other whodunit cases, if Perplexity is off the hook, where does that leave Perplexity’s users legally? Could Amazon assert CFAA/502 claims against users, or a range of other legal theories? In a footnote, the court says tersely: “This outcome does not impair Amazon’s ability to regulate access to Amazon.com via private terms of service for its users.”

The users’ legal exposure highlights some challenging policy fault lines of this case–and all online trespass to chattels cases. On the one hand, it would be a bad thing if websites more aggressively erect contract-based putative barriers to access. That creates a potential thicket for ordinary daily activities, and it raises greater risks that individuals will unintentionally trigger significant legal exposure if they don’t research and confirm their permission to visit every service they visit.

On the other hand, Amazon clearly expressed its preferences that it didn’t want Perplexity’s users to access its services through agentic AI, and why shouldn’t the law back up these strong and clear desires? We know that server operators need that prerogative to fend off hostile attacks, but we also know server operators will overclaim any legally backed prerogative they have.

As a result, as with most trespass-to-chattels cases, there are not likely to be any easy answers in this case. There are only significant and potentially problematic policy tradeoffs.

The post Ninth Circuit Lifts Restrictions on Agentic AI Accessing Amazon appeared first on Technology & Marketing Law Blog.