New Class Action Accuses Instagram of Mass Copyright Infringement

On May 19, two California photographers filed a class action suit alleging that Instagram is liable for inducement of copyright infringement, contributory copyright infringement, and vicarious copyright infringement due to the platform’s implementation and promotion of its “embed” function. The complaint begins: “This case is about Instagram’s scheme to generate substantial revenue for its parent, Facebook, Inc., by encouraging, inducing, and facilitating third parties to commit widespread copyright infringement.” Assuming this case proceeds, it could be a doozy, but first a little background.

The Server Rule and Terms of Service

Simply put, when an Instagram user uploads an image to a public account, a third party can cause that image to be displayed on its own web pages by means of copying HTML code that “points” to the image on Instagram’s server without copying the image onto the third party’s server. This act of “pointing” to a file on a host server is referred to as embedding or in-line linking or sometimes just linking.

Since 2007, and until quite recently, the practice of embedding, without making a copy on a different server, was held not to infringe copyright, thereby establishing the judge-made “server rule” in the case Perfect 10 v. Amazon. But more recent jurisprudence, particularly in instances when third-party users are commercial enterprises, courts have recognized that although embedding does not implicate the reproduction right, it still implicates the display right, which is simply common sense to most of us.

If a news organization causes a photograph to be shown on its web page along with a story it publishes, this is unquestionably a “public display” of the work. And the benefit derived by the news organization’s use of the image is a paradigmatic example of a use that requires a license. And that is a digest version of what a New York district court held in 2018 in Goldman v. Breitbart et al, after which several news organizations settled with photographer Justin Goldman for unlicensed display of one of his photographs by means of embedding from Goldman’s Twitter feed.

Thus, after Goldman, the server rule was held (and I think correctly) not to relieve the end-user third party from its obligation to license a work because the use still implicates the exclusive right of public display. At least in the SDNY. Then, in a related case (Sinclair v. Ziff-Davis), photographer Stephanie Sinclair complained that Mashable had caused her unlicensed image to be displayed, via Instagram’s embed API, for one of its articles.

But there, a New York district court articulated nothing about the server rule (or copyright per se) because the judge stated that she had no choice other than to affirm that any user who uploads an image to a public Instagram account grants that platform a very broad license to sub-license works to third parties. As discussed in my post about this ruling, under its TOS, Instagram can even charge fees to third parties for these sub-licenses, but so far, the platforms have not been quite so brazen.

Partly in response to Sinclair, a spokesperson for parent company Facebook stated to Ars Technica, “While our terms allow us to grant a sub-license, we do not grant one for our embeds API.” Ars also reports in the same article (albeit under the banner of whining about “web culture”) that another judge weighing a similar case in the same district (McGucken v. Newsweek) had not been so quick to dismiss the complaint on the TOS-based precedent articulated in Sinclair.

Holding Instagram Liable

None of those cases, of course, implicates the source platforms where the copyright owners originally upload their images. And that brings us to the class action Hunley et al v. Instagram, which alleges that the Facebook-owned platform has intentionally caused chronic third-party infringements of primarily photographic and videographic works. To meet the standards under the allegations filed, plaintiffs will have to demonstrate, to put it in simple terms, that Instagram’s managers knowingly fostered third-party infringements and did so for its own material gain.

But before this case enters that discovery phase, I imagine Instagram will file a motion to dismiss based on both the server rule and the TOS contract to which the plaintiff class has agreed by using Instagram in the first place. In the first matter, it may prove unfortunate for the plaintiffs that this complaint was filed in the Ninth Circuit, where the server rule is precedent, while it is only in the Second Circuit where district courts have articulated that the rule should not extend to infringing the display right. If the California district court defers to the server rule, it may find that the third parties did not infringe and, therefore, no basis for secondary forms of infringement exists. Alternatively, in a potential win for rightsholders overall, courts in the Ninth Circuit will agree with the display right limit, and at least the server rule will cease to be a free pass for parties to infringe visual works.

On the TOS question, the plaintiff class could, at least initially, run into the same problem as Sinclair because the language granting Instagram the right to sub-license is quite clear. Just because a Facebook PR person stated that third parties need to obtain licenses, that does not amend the contract with users (TOS) giving Instagram the authority to grant sub-licenses. The plaintiff class will seek to prove that Instagram is playing games—that although it is not exercising its contractual authority to sub-license, it has instead engaged in conduct to induce third parties to intentionally and chronically infringe works, and it has done so for its own material gain. From the complaint:

Embedding was originally marketed as a new tool to share one’s own photos or videos from one’s Instagram account to ‘your own website or blog.’ But Instagram had other plans that it was less public about, including the embedding of copyrighted works onto third party websites and blogs that were hosted outside of Instagram’s platform. That month [July 2013], Instagram rolled out a ‘pilot program’ with Bleacher Report, Mashable.com. CNN.com, HuffPost.com, and People.com to test drive the new ‘embedding’ tool.

To secretly enhance Instagram’s embedding tool’s functionality and generation of revenue, in 2016 or earlier, Instagram started putting tracking code technology that attached to users’ photos and videos uploaded to Instagram as a means to measure, handle, monitor, track and monetize the most valuable copyrighted works, i.e. the works that generated the most traffic.

Certainly, these allegations (and others in the complaint) are consistent with the manner in which platforms exploit the traffic or data generated by third-party infringements, while claiming to have no involvement in the infringements themselves. As such, especially if the plaintiff class can overcome the server rule hurdle in the Ninth Circuit, this case could prove to be a game-changer for online copyright enforcement. On the other hand, a class action against a giant like Facebook could last a decade or more, at which point the proverbial game may look very different.

Podcast Interview “Lost in the Stacks”

Had a fun conversation with Fred Rascoe of Lost in the Stacks: the Research Librarian Rock ‘n’ Roll Radio Show, out of Georgia Tech University, broadcasting on WREK Atlanta. We talked about copyright, creativity, and my book Who Invented Oscar Wilde?

A Response to “Weaponizing Copyright” by Cathay Y. N. Smith

Last week, I stuck my toe into a little debate on Twitter about the word weaponize, when Professor Cathay Y. N. Smith[1] defended her use of the expression “Weaponizing Copyright.” Smith was citing the title and subject of her own draft paper, and because I still hate Twitter for discussing complex issues, I read the 73-page draft over the weekend. While I disagree with several important aspects of Professor Smith’s approach, it is a thoughtfully written paper so far, highlighting more than one point of view, and despite the choice of a loaded term like weaponize, it is not as thoroughly hostile to copyright as its title suggests.

Professor Smith’s main thesis is that the use of copyright enforcement for “noncopyright purposes,” which is how she defines “weaponizing copyright,” is increasing, though she does not contend that all alleged “weaponization” of copyright is necessarily a bad thing. She identifies, “five common noncopyright objectives that copyright owners seek to achieve through weaponization: to silence and erase facts, suppress criticism and speech, punish and retaliate, protect reputation and moral rights, and preserve privacy.” For each of these five categories, Smith cites various examples, some of which I would argue do not belong in her data set, even as anecdotal evidence. But I’ll get to those in a moment.

Defining the Purpose of Copyright

The premise of Weaponizing Copyright begs some fundamental questions that are either not answered or not answered correctly in my view. First, how are we defining copyright’s purpose? Second, is “weaponizing” an appropriate term for noncopyright enforcement action? And third, is copyright “weaponization,” even if we accept the term, truly increasing? And perhaps a subordinate of question three is to ask how successful are “weaponized” copyright enforcement actions?

Smith does not answer the latter questions at all. Her paper describes several enforcement examples, but even if all these examples properly fit the definition of “noncopyright objectives,” a handful of anecdotes is not a sufficient body of data to show whether the types of legal actions described are increasing. But even before that, I would argue that not every example cited serves noncopyright ends because Smith begins with a too-narrow definition of copyright’s purpose, one which elides non-pecuniary (or non-market) reasons for enforcement. She writes, “Copyright grants authors exclusive rights in their works, which allows authors to realize financial and economic gains for their works.” True but incomplete.

It is the view of many copyright skeptics that copyright should never serve any purpose other than to protect an author’s financial interest in her works, but this opinion is inconsistent with history, case law, and a basic sense of fairness. As Professor Jane Ginsburg observes in a paper from 1990, both the natural rights (French) and utilitarian (English) rationales for copyright were alive at America’s founding period.[2] And without going into a long history discussion, suffice to say that non-pecuniary reasons for protecting authorial rights have been part of our philosophical consciousness since before modern copyright existed, and these rationales are still part of the copyright narrative. Sometimes, we simply call it justice.

Non-Financial is Not Noncopyright

Because copyright protects works of creative expression, the exclusive right of the author to license the use of a work naturally anticipates that the author might reject certain uses, regardless of financial consideration. I doubt very much, for instance, that Ta-Nehisi Coates would allow the exploitation of his literary works by the KKK, even if they offer him millions of dollars. And if such a group infringed his works, I imagine he would file suit for the same reasons that he would have rejected a license in the first place. This non-pecuniary motive is not in any way inconsistent with the purpose of copyright and, therefore, should not be considered an example of “weaponizing” copyright, if we are to use that term at all.

So, although I agree that Professor Smith alludes to some noncopyright examples of enforcement, I would argue that her incomplete definition of copyright’s purpose encompasses other examples that properly manifest the purpose of copyright. For instance, Matt Furie’s attempt to litigate against white supremacist and fascist groups for mass appropriation of his Pepe the Frog character should not be included in Smith’s discussion in my view. Furie’s complaint is wholly consistent with the purpose of copyright, responding to appropriations that obliterated the value of his work in every sense. Pepe was simultaneously damaged as a work of expression and as a work of potential market value, and it is copyright’s purpose to protect both.

Likewise, Professor Smith should not include William Greenblatt’s efforts to stop the gun wielding McCloskeys from using his photograph as a holiday card. Although Smith acknowledges that Greenblatt might have several motives and she is, therefore, hesitant to call his enforcement “weaponization,” none of that matters. When an author says Do not reproduce my photo as your holiday card, this demand fits squarely within the purpose of copyright, and Greenblatt’s motives, while we can probably guess what they are, are irrelevant to a discussion about noncopyright purposes.

Where I am more intrigued by Professor Smith’s inquiry is when she cites an example like The Weinstein Company attempting to use copyright law to stop Ronan Farrow from exposing evidence to support the sexual assault allegations against Harvey Weinstein. I think we all agree that trying to stifle news of compelling public interest is well beyond the purpose of copyright, and in the Weinstein example—a bare-knuckle fight with a bully—it seems appropriate to use a term like “weaponize.”

On the other hand, Weinstein’s attempt to frighten Farrow with the specter of copyright litigation was not successful, which brings the conversation back to the need for a broader data set, if we are to consider whether Professor Smith is truly examining a subject that needs addressing or citing anomalies that are simply unavoidable in legal battles. No doubt, just about every area of law has been misused in some way by someone, but that does not mean we necessarily amend or scorn the body of law itself.

Related to this need for more robust data, when Professor Smith cites examples adjacent to Weinstein, in which celebrities seek to protect reputation and/or privacy by enforcing copyrights in images, video clips, or even emails and other written material, I believe the conversation becomes more nuanced, and I would again question whether “weaponize” is a proper term. Almost everyone would agree that the public has a compelling interest to know the facts about a business mogul who may be a sexual predator, a fraud, a drug abuser, etc., but the presumed “right to know” probably does, and certainly should, diversify our views when it comes to the merely gossipy and prurient.

For instance, Smith is right that it is not very close to the soul of copyright to enable Hulk Hogan to sue for the unauthorized release of his sex tapes, but neither is it generally beyond the purpose of copyright to protect the author’s right to not publish a work. Granted, the word author in regard to a sex tape is a bit much, but the legal principle is the same. Copyright enforcement is the remedy when unpublished works are made available without license. And again, it does not matter whether the works are unpublished because the author would be embarrassed by them or because she decides not to publish for any other reason.

Correspondingly, does the public really have the right to see Hulk Hogan’s, or anyone else’s, sex tape? Or some celebrity’s secret wedding photos? If a person of influence has a sex tape (or gets married), perhaps the public has a right to know the facts (because the facts may be relevant to the subject’s public role), but that does not extend to a right to see the works per se. Samuel Warren and Louis Brandeis, who wrote the seminal paper on privacy in American law, would agree with this, and when they began seeking a foundation for privacy, they turned first to copyright law and the protection of unpublished works.[3] So, once again, I would personally disqualify some of Smith’s examples as standing far outside the penumbra of copyright’s purpose and, therefore, deserving the term “weaponize.”

As stated above, Professor Smith does not necessarily reject every noncopyright purpose of enforcement as a negative use of copyright. For instance, she entertains the idea that perhaps copyright should serve to protect privacy. But my point is that there is an extent to which copyright always has protected privacy with the protection of unpublished works, so it is not necessarily a novel principle, even if the digital age has provided so many new opportunities for privacy to be invaded. At the same time, I think Smith poses a thoughtful question when she writes the following:

… if we are content with copyright being used for purposes of protecting privacy, reputational or dignitary interests, could the law become complacent in holding back important and necessary reforms of laws, such as in the areas of privacy, Section 230, and others?

I believe many legal authorities, including copyright advocates, would prefer to see, for instance, Section 230 no longer protect websites that trade in nonconsensual pornography, rather than have victims use DMCA takedown as a half-measure to remedy this form of harassment. Copyright enforcement is a workaround—a partial solution to a very ugly crime, and DMCA takedown is inadequate relative to the kind of damage inflicted or the motives for this crime to be committed in the first place. That said, I very much doubt the copyright stopgap solution will stand in the way of improving other areas of law, where improvement is possible. We seem to be making progress toward a federal prohibition against nonconsensual distribution of intimate images, and I don’t believe anyone has seriously suggested that DMCA takedowns obviate the need for this legislation.

On the subject of the DMCA, I think Professor Smith accurately describes certain uses of the takedown provision on platforms like YouTube to occasionally silence criticism or speech, or to punish. And she is also not wrong that DMCA takedown can be a relatively simple means to achieve these ends. But, on a side note, I also believe that inquiries in this area should compartmentalize sites like YouTube due to the intramural nature of copyright enforcement on the platform. Meritorious claims are in the mosh pit with unmeritorious claims, mixing it up with the ever-changing motives of YouTubers, and all of this chaos is administered by a corporate landlord with its own systems, interests, and dynamic rules while it adjudicates both takedown notices and counter-notices. YouTube is its own special copyright circus.

But more acutely, one of Smith’s key YouTube examples, citing gamer/YouTuber PewDiePie, reveals once again that she is working with an incomplete definition of the purpose of copyright. Once the highest paid YouTuber in the world, PewDiePie gained internet stardom with videos of himself playing popular games while joking, critiquing, or commenting as he played. In general, the game developers saw this use of their materials as good marketing, but in 2017, PewDiePie’s star fizzled a bit when he made racist and sexist remarks in several of his videos. Suddenly, “Pewds” was no longer a positive association for every game developer.

Professor Smith describes game developer Campo Santo’s takedowns as serving the noncopyright purpose of seeking to “punish” PewDiePie for being a jerk. But this is an error in my view. Because PewDiePie’s uses of gaming material likely would not be protected by fair use,[4] any of the gaming companies’ decisions not to issue takedowns should be understood as unwritten licenses to allow PewDiePie to use their works. But these are conditional licenses. As long as the use is good for the brand, the use is allowed. If the user is no longer good for the brand (just like an actor on a TV show whose public words or deeds become a liability), the company may revoke its unwritten license.

In other words, once again, there is nothing tangential to copyright’s purpose in Campo Santo’s decision to file the takedown notices—even if it turns out to be a financially detrimental decision, as Smith indicates it might be. Rightsholders grant unwritten, conditional licenses all the time (especially on the internet), simply by allowing uses they discover, and without any kind of formal agreements. But if the conditions of a particular use change, the copyright owner reserves the right to revoke that unspoken license. In fact, this is categorically a purpose of copyright.

I cannot address every example or aspect of Professor Smith’s paper in a single post, and my critiques here do not deny the fact that, in powerful hands, even the threat of copyright enforcement can be used as a weapon against a less powerful individual or entity for purposes outside the spirit of copyright’s intent. Smith is not wrong when she describes the attributes of copyright law that make it well suited for certain parties to try to silence criticism or erase information or perhaps even punish. And I think she is particularly thoughtful in exploring how and whether copyright is properly used to defend against powerful aggressors, even in contexts not typically thought of as fulfilling the spirit of copyright.

But recognizing that just about any legal framework may be used in a way we’ll call unconventional, questions remain as to how often that applies to copyright, and how often owners truly achieve noncopyright ends, regardless of whether those ends are ill or good. Meanwhile, before such an inquiry can provide useful insight, the examples considered should be based on a more complete definition of the purpose of copyright, and consequently a much narrower definition of what it means to weaponize.


[1] Associate Professor of Law, University of Montana Blewett School of Law.

[2] https://scholarship.law.columbia.edu/faculty_scholarship/620/

[3] https://www.cs.cornell.edu/~shmat/courses/cs5436/warren-brandeis.pdf

[4] Using portions of a video game in a video revie would likely be fair use, but using hours of game play as a major component In your for-profit entertainment vehicle would disfavor a finding of fair use.


Puzzle source image by: AndreyPopov