In-Line Linking May Be Copyright Infringement–Goldman v. Breitbart News
Ugh, this decision is bad. How bad is it? It makes me sympathetic to Breitbart, and I didn’t even know that was possible. You may want a box of tissues nearby before reading this. The TL;DR: for over a decade, in-line linking has been treated as categorically non-infringing. This opinion flips that presumption and may eliminate all unlicensed in-line linking.
Goldman (no relation) posted a photo of quarterback Tom Brady to Snapchat. It went “viral,” and third parties reposted the photo to Twitter. The defendants then “embedded” those tweets, including Goldman’s photo. Functionally, embedding is the same as in-line linking. The photo remains hosted on Twitter’s servers, and embedding automatically instructs web browsers how to obtain the photo from Twitter’s servers and incorporate it into the page’s HTML.
The parties agreed to put a threshold question to the judge first: is embedding tweets a “public display” for copyright purposes? All other legal questions were reserved for subsequent proceedings depending on the court’s answer. Because of the streamlined legal question answered by the court, any gossipy details about Brady and Twitter are irrelevant. This ruling is much broader than that.
The court says in-line linking of photos constitutes a prima facie copyright infringement (a public display). The defendants can still advance other defenses, and they might even win on other grounds. But even if that happens, this ruling will remain deeply troubling if it’s not fixed on appeal.
To reach its shocking conclusion, the court navigates around the decisions of two other circuits (and several other district court cases). Most importantly, the Ninth Circuit’s 2007 Perfect 10 v. Amazon opinion adopted a”server test,” and rejected the “incorporation test,” for public display. The “server test” means that in-line linking constitutes a public display only when the defendant hosts the photo on its server. If the defendant uses other technical means to show a photo without hosting it, it’s not a public display and therefore categorically not copyright infringement. The Seventh Circuit’s 2012 Flava Works v. Gunter opinion adopted a similar conclusion for showing video. These two rulings, dating back over a decade and with minimal conflicting precedent, led to many settled expectations and the proliferation of in-line linking.
To distinguish the adverse precedent, the court says:
The plain language of the Copyright Act, the legislative history undergirding its enactment, and subsequent Supreme Court jurisprudence provide no basis for a rule that allows the physical location or possession of an image to determine who may or may not have “displayed” a work within the meaning of the Copyright Act. Moreover, the Court agrees that there are critical factual distinctions between Perfect 10 and this case such that, even if the Second Circuit were to find the Server Test consistent with the Copyright Act, it would be inapplicable here.
The court says that the statute defines “public display” broadly, and “each and every defendant itself took active steps to put a process in place that resulted in a transmission of the photos so that they could be visibly shown.” (Yes, that’s what it means to in-line link). Citing Aereo as persuasive authority, the court indicates that form matters more than substance: “liability should not hinge on invisible, technical processes imperceptible to the viewer.” Aereo’s invocation isn’t surprising: when I recapped Aereo, I raised the viability of in-line linking as one of four questions left unanswered by the opinion. Still, when the Copyright Act bases so many legal rules on technical minutiae, it’s extremely aggravating for courts to say the technical details don’t matter.
The court expressly rejects Perfect 10 because retailers must pay to play the radio in their stores, which proves that possession of a “copy” isn’t required for copyright infringement. The court also says that the Perfect 10’s server test was specific to Google’s search engine functionality and based on Google’s implementation requiring users to click on a link before they could see the in-line link.
The defendants argued that the “incorporation test” would “cause a tremendous chilling effect on the core functionality of the web,” and amici argued it would “radically change linking practices, and thereby transform the Internet as we know it.” The court responds that the defendants in this case might still win in further proceedings. That’s true, but it’s a tone-deaf response because the court’s ruling undermines the situation for everyone else. The court has eliminated a bright-line rule that many Internet actors rely upon, and having to rely on other defenses will increase defendants’ costs and reduce their certainty.
This ruling reaches quite far. Let’s start with Twitter. According to the court, everyone using Twitter’s embed function is committing prima facie copyright infringement. Twitter can partially address that problem through a TOS license that authorizes third party usage of the embed feature. However, Twitter’s TOS can’t fix the problem that users upload infringing items to Twitter, which by definition the TOS can’t authorize for relicense, so showing the infringing images through the embed functionality will lead to strict liability prima facie copyright infringement. In other words, a single infringing upload by a Twitter user potentially virally contaminates everyone else–potentially thousands of people–unwittingly using the embed feature, exposing all of them to financially crippling copyright litigation. I could see why Twitter might rationally decide to turn off the embed function entirely, because otherwise the tool facilitates copyright infringement by users who don’t know better.
Furthermore, in-line linking has become quite ubiquitous, especially in social media where it’s common to provide a preview of an article/page that a user links to. These previews are typically in-line linked, but the entire practice is now risky. There’s nothing these social media sites can do to mitigate that risk through their TOSes. Similarly, all image search engines may feel like they need to turn off in-line linking now (though Google apparently just did that as part of its settlement with Getty), and there is nothing they could do by contract to fix this risk.
Because of the widespread mischief created by this ruling, I hope the defendants appeal it to the Second Circuit. Unfortunately, the appeal will be risky. The Second Circuit has issued some favorable defense-side online copyright rulings (e.g., Cablevision, the Google Books fair use ruling, Vimeo), but it’s also issued some problematic clunkers (e.g., Viacom v. YouTube, UMG v. MP3Tunes/Escape Media). If the Second Circuit affirms the ruling, there may be enough conflict with the Ninth and Seventh Circuits to support Supreme Court certiorari. However, the defendants might choose to settle the case or may win it on further proceedings, leaving this ruling hanging without an appeal. And the longer this opinion remains on the books, the more trouble it will cause.
Case citation: Goldman v. Breitbart News Network, LLC, 2018 WL 911340 (S.D.N.Y. Feb. 15, 2018).