[Disclaimer: Because this article is intended to be both instructive to the general public but also legally relevant, in the course of my splitting the difference and trying to serve two audiences, the general public may find some of it legalistic and dense, while lawyers may find some of it pedantic. Apologies in advance to whichever camp you find yourself in, and feel free to contact me [scott.pilutik[AT]tikk.net] if you have any questions.]
I’ve been following with some interest the upstart social network Pinterest, at first merely to gain some insight into its out-of-the-blue popularity. This wasn’t easy because I sit manifestly outside Pinterest’s demographic as a 43-year old man with an interest in news, politics, sports, and the law. The default Pinterest landing page appeals to none of those things. Rather, it’s an online scrapbook curated by women aged 25-45 who evidently read Crate & Barrel and J. Crew catalogs along with Martha Stewart magazines.
Pinterest provides a simple interface to collect and share home decor, craft projects, fashion ideas, and the like. Users can categorize their “pins” onto boards–e.g., “Great Food,” “Fashion,” and “My Style.” Under the hood, though, Pinterest is content–and gender–neutral and operates rather simply. One “pins” by pasting a URL (which can be a web page or an image file, but there must be an image on the page), typing a description, and selecting, or creating, a category. That’s it.
Finally struck by this epiphany that Pinterest was merely a well-designed organizational tool, I unfollowed nearly all my Facebook friends and family (naturally, mostly women) whose pins dominated my feed and began using Pinterest to bookmark articles about what interested me most about Pinterest–its many-faceted Copyright Problem.
The copyright problem is easy to spot if you’re an intellectual property attorney (on my Facebook page I called Pinterest an “infringement farm” ten minutes after my first visit), but especially glaring if you’re an attorney who is also a photographer. Like myself, Kirsten Kowalski qualifies as both, and on February 24, 2012, expressed her concerns in a lengthy blog post titled Why I Tearfully Deleted My Pinterest Inspiration Boards.
Pinterest Terms of Service and Your Liability
Kowalski plunges into both Pinterest’s legalese and federal law to reconcile the disconnect she saw between Pinterest’s implicit encouragement to propagate third parties’ copyrighted works and its terms of service, where it requires each user to “represent and warrant” that they possess (either as owner or licensee) the republishing rights of the works they Pin; and that their use on Pinterest violates no other third party’s intellectual property rights.
There is nothing especially atypical about Pinterest’s terms of service in this regard: “Online Service Providers” (OSPs, defined by US copyright law), such as Pinterest, only initially assume liability for the content users post to their sites. The Digital Millennium Copyright Act (DMCA) contains a safe harbor mechanism permitting Pinterest to effectively indemnify itself from copyright infringement lawsuits by timely removing alleged infringements upon notification. The process then permits users to “counternotify” after which time the OSP (Pinterest) will restore the image without fear of lawsuit, which risk is now solely borne by the user in their personal capacity.
In other words, the terms of service language about which Kowalski is most concerned isn’t especially different than most other sites’ terms of service, which, in turn, mostly restate the law with respect to assumption of liability for infringing conduct. [Note: As I was writing this post Pinterest removed from its terms of service some of its more egregious provisions, such as the right to sell your work, and a perpetual, irrevocable sublicense to your work. These and other changes made as of March 24, 2012, do little to alter the issue with which this post is chiefly concerned, namely liability for copyright infringements.]
Kowalski’s concerns aren’t unwarranted, though, because the manner in which Pinterest is commonly used by its users makes it inevitable that thousands of copyright infringements will be committed each day. But Kowalski’s fear of an infringement lawsuit has less to do with Pinterest’s terms of service, which merely reflect existing law, than with Pinterest’s business model, which relies on its users republishing unlicensed third party content. Pinterest’s terms of service and its business model are at odds, and may even serve to confuse its users, though it seems unlikely that such conflict would shift liability for infringements from Pinterest’s users back to Pinterest.
If Pinterest’s users do indeed risk an infringement lawsuit with each pin, it is worth examining first the likelihood of that risk, and then such a lawsuit’s chances of success, while also considering the possible defenses available to a Pinterest user sued for a typical infringement.
Pinterest Users — Legal vs. Real Risk
Kowalski doesn’t tackle the risk question, except to briefly allude to the fear that Pinterest is the next Napster. There are good reasons, though, both real-world and legal, why a hypothetical Pinterest user/defendant is distinguishable from the Napster defendants. The subject matter of a Pinterest-related “infringement” is an image, not a music or video file, and there are no photography rights-holding trade organizations on the scale of the RIAA and MPAA, who both can allocate significant resources to policing infringement [though many relatively small photo trade organizations exist; see here].
Because there are significant disincentives for photographers to sue for web infringements–e.g., the expense of bringing an infringement suit; obtaining personal jurisdiction over out-of-state and possibly judgment-proof defendants; nominal infringements can mean nominal damages– the cost of bringing such a suit easily can and most often is offset by the potential return. And without the luxury of RIAA/MPAA-sized resources, these cost-based risks fall on the photographer/plaintiff.
A Pinterest-based infringement is also distinguishable from Napster legally, largely because there is little room to lodge a fair-use defense for audio and video file-sharing infringements, as each audio/video file most often represents the entire infringed-upon work, untransformed in any way, and which is usually available for sale online.
So while an infringement suit against you as a Pinterest user is unlikely, a suit against some Pinterest user seems inevitable, and its outcome may serve to shape the future of Pinterest and other scrapbooking social media sites.
The Fair Use Defense — Introduction
For the unlucky Pinterest user who finds him or herself a defendant, what is their recourse? Presuming such lawsuit is adequately pled with respect to matters like jurisdiction, a defendant would argue that their use was fair, pursuant to Section 107 of the Copyright Act. Fair use can broadly be described as copyright law’s acknowledgment that certain per se infringements should nevertheless be permitted in the interest of fairness, education, and economy (not to mention in the interest the promotion of the progress of science and art, pursuant to Art. 1, Sec. 8 of the U.S. Constitution, which is a too-broad discussion for the moment).
Analytically, whether use of a work is fair is tough to determine and predict. Per Section 107, a nonexclusive four-factor test considers each of the following: the (1) purpose and character of your use; (2) nature of the copyrighted work; (3) amount and substantiality of the portion taken, and (4) effect of the use upon the potential market. I’ll refer back to these factors by their number later in this post.
No one factor is necessarily determinative, and the myriad fair use cases are of limited help, as different courts can fixate on different factual distinctions, together with the potential uniqueness of each scenario underscoring the lawsuit. These difficulties are compounded by the difficulty both attorneys and courts have had applying old copyright law to new technologies, which are often misunderstood by these new technologies.
Fair Use First Factor — Nature of Use
That disclaimer aside, some fair use cases are possibly on point. Kowalski raises an obvious one, the well-known Kelly v. Arriba Soft Corp., 336 F.3d 811(9th Cir. 2003) case, in which a photographer sued an image search engine (prior in time but similar to Google’s image search) for republishing thumbnail versions of the photographer’s images, the larger version which the photographer had himself hosted.
The Ninth Circuit analyzed each of the above-mentioned four factors in turn, concluding that Factor 2 weighed in favor of the photographer (reasoning that creative works merit stronger protection than non-creative works, but already-published works merited weaker protection); Factor 3 weighed in favor of neither party, the court reasoning that even though each copyrighted work was copied in its entirety, the search engine only rendered as much as necessary (if this sounds tautological to you, you’re not alone); and Factor 4 weighed in favor of the search engine, because far from harming the market for the images, the search engine aided the market by pushing traffic back to the photographer’s site.
The court discussed Factor 1 in significantly greater depth than Factors 2, 3, and 4, finding in favor of the search engine that the nature of the use was “transformative” inasmuch as the purpose of the use differed entirely from the plaintiff photographer’s use. A search engine republishing thumbnail images involves “more than the retransmission of the images [into] a different medium.” A necessary recontextualization takes place in which the altered image not only serves a new purpose which provides a public benefit (by way of a private benefit to Arriba Soft, of course). The court also remarked on the thumbnail image’s lack of further utility, because enlargement could not occur without sacrificing clarity.
As Kowalski rightly points out, the Arriba Soft case does not literally stand for the proposition that thumbnail images are uniformly fair use exceptions, because no one fair use factor is determinative. But the reality is that image search engines such as Arriba Soft (now Ditto.com) and Google Images would not exist if there were any legal uncertainty as to whether thumbnail images of existing copyrighted works qualified as fair use. They do, period.
A few years after deciding Arriba Soft, the Ninth Circuit found similarly in an almost identical case, Perfect 10 v. Google [captioned as Perfect 10, Inc. v. Amazon.com, Inc., 487 F.3d 701 (9th Cir. 2007)]. Perfect 10 sued Google for “in-line linking” to its copyrighted images and rendering those images as thumbnails on its search engine. The Court followed the same fair use analysis as in Arriba Soft, and was further unpersuaded that the technical distinctions between Google’s and Arriba Soft’s search engines changed the outcome (unlike Arriba Soft’s, Google’s search engine presents a thumbnail of the actual copyrighted work within its own frame).
Applying Arriba Soft & Perfect 10’s First Factor Analysis to Pinterest
Presuming a hypothetical case against a Pinterest user (I’ll discuss Pinterest’s liability as a secondary infringer separately) is brought in court under the Ninth Circuit, it is likely that Factors 2, 3, and 4 would be analyzed similarly. Arriba Soft and Perfect 10 are only persuasive authority within the Ninth Circuit, though both have been cited widely and favorably by other federal courts.
However, Factor 1, the purpose and character of the use, could be a different story. In Arriba Soft and Perfect 10, the court was only confronted with search engines having rendered copyrighted works as thumbnails. Pinterest generates at least two, and likely more, copies of each image associated with every user pin, and at least one of those images is much larger than a thumbnail.
The first is a thumbnail approximately 192 pixels wide, which is comparable in size to the thumbnails generated by Google and Arriba Soft [now Ditto], and are the most common image viewed by Pinterest users, appearing on the user’s main feed showing the activity of the users they follow.
From this thumbnail-filled landing page, users can click on theirs and their followers’ thumbnails, which takes the user to a page containing a larger image of the same thumbnail, the description for the image and user comments, if any. This image can be at least 550 pixels wide (per my rough, unscientific survey). Clicking on this large image takes the user off site to the source URL the user inputted when creating the pin, which, again, can either be an image file or a web page containing an image.
As the above makes clear, Pinterest automatically generates both a thumbnail and a large copy of each source image, which it automatically republishes and presents. The size of this larger copy may even identical to the source image size, if the source image size is within Pinterest’s approximately 550 pixel wide layout constraint.
To see how little a source image can be transformed from source to Pinterest, start with the following Pin of a young boy seated on the lap of an older man waring a cowboy hat.
Clicking on the Pin takes the user to a web page which contains the same image displayed at the Pin:
The image file seen at the shorpy.com page is 439 px wide & 290 px tall, as seen here:
The copy of the above image generated by Pinterest is 439 px wide & 290 px tall, identical to the source image in every sense, as seen here:
That thumbnail images can easily be distinguished from exact copies is not necessarily an outcome-determinative distinction, though. As indicated above, the four fair use factors are applied in a highly unscientific “totality of circumstances” manner, making it difficult for parties to rely on factual similarities and distinctions between decided cases and your own. But to the extent that Arriba Soft and Perfect 10 have come to stand for the proposition that “thumbmail images are fair use,” it is difficult to argue their applicability to Pinterest, whose users create both thumbnail and exact copies of hundreds of thousands of images, the overwhelming majority of which are posted without their rights ever having been negotiated.
A fair argument, however, can nevertheless be made that the purpose and character of the use (Factor One) is substantially transformed even if the use includes the republishing of an exact copy. As can be seen in the above screenshot, the copied image, when placed within Pinterest’s environment, is completely recontextualized. The user’s name who pinned the image/site is identified above the image, along with the time he/she posted it. Below the image is the user-provided caption, and a form field to enter comments, above any comments that might already have been left. Below the comments are relevant meta-data, such as the identity of the user who originally pinned the image/site, the name of board (user-determined category) onto which it was pinned, the technological mechanism by which it was pinned (e.g., a bookmarklet, web), etc.
Accordingly, it may be argued that like Arriba Soft and Google, the purpose and character of each image’s use is functionally distinct from the rights holder’s purpose and character of use. A Pinterest user effectively curates collections of imagery organized and defined by his/her own parameters, so not only is the appropriated image surrounded by a new description and new meta-data, but it is further recontextualized by how the user categorizes it–what other images it’s surrounded by. Further, the more commentary that is added (including by the user’s followers), the more the image is additionally recontextualized, and the more fair the use, or so goes the contention.
This is, in essence, the argument recently put forward by intellectual property attorney Itai Maytal, who was quoted by Business Insider asserting that Pinterest’s requirement that users add a caption to each Pin lends itself to a favorable fair use defense because the caption is “commentary, and commentary and parody are some of the types of uses that the law encourages[.]” Maytal concedes that the fair use inquiry would not end there, but even so disclaimed, the argument is problematic.
It has never been held, to the extent I can find, that an exact, unaltered, uncropped copy of an image can be sufficiently transformed, for the purpose of a fair use defense, merely by recaptioning, or adding a caption to it. It’s hard to imagine, for example, the New York Post republishing an image it copied from the New York Times website and then suggesting it had every right to do so because it ran a caption whereas the Times did not (or ran a different caption). An apt parallel media comparison might be a broadcaster attempting to argue that its disc jockeys’ talking before and after songs sufficiently transforms the songs so as to moot the need to pay royalties. If only effecting transformative use upon copyrighted works was as easy as sprinkling a few words around its edges, it would be too easy to circumvent and abuse of such a lazy standard.
There are scenarios in which captioning/recaptioning an otherwise uncropped, non-thumbnail image might tend to qualify as fair use, such as in the context of an academic paper, where the image itself were the subject of scholarly analysis, or was the subject of a parody. By contrast, Pinterest users are typically not republishing images to parodic or scholarly ends, but rather to assert those images place within the user’s particular taste spectrum. Users are essentially rogue curators, republishing unlicensed images as a personal, exhibitionistic exercise, proclaiming to their followers, “These images are me.” The amount of creativity that goes into a scholarly or parodic work of the sort held out as typical fair use examples, and a Pinterest caption is stark indeed.
Fair Use Factor Four - Marketplace Harm: Law & Economics
Though, as I observe above, Pinterest’s users appear vulnerable with respect to the first Fair Use factor, the fourth (again, non determinable) Factor, which disfavors uses harmful to the copyrighted work’s potential market, holds some promise. Pinterest’s users typically do not seek to profit from republishing unlicensed works, and, more important, rights holders typically do not suffer from Pinterest’s users’ republication of their works. Indeed, because Pinterest requires each Pin to link back to its source material, Pinterest is, if anything, improving the potential market of unlicensed works by exposing its users to them.
Law professor Dave Fagundes, writing at PrawfsBlawg, makes the argument that a plaintiff suing a Pinterest user would face an uphill climb trying to prove that the republication of their image harmed its marketability. This is because in order to argue that the potential market is being harmed, a rights holder must show both that a real market exists and that the works’ republication on Pinterest consequently deprives that market.
As an ex-commercial photographer I’m at least nominally qualified to speculate on the types of markets that unfettered infringements by Pinterest users conceivably might harm. With regard to photographic prints, it might be argued that online copies can never serve as an adequate market substitute for physical copies unless such copies are of a high enough resolution to enable the printing of art-quality versions. But Pinterest’s largest version easily fits within the four borders of any typical computer screen and pose no threat to become repurposed as works of art, print or magazine advertisements, or to useful beyond the digital web-only environment.
Photographers who derive revenue from stock photography agencies could potentially be adversely affected by users’ republishing their unwatermarked images on Pinterest. Stock images are generally sold as hi-resolution files, far greater in size than Pinterest’s largest rendered size, but not all purchasers of stock imagery need the highest-resolution file–for web use, the Pinterest-published version of an unlicensed stock image may serve as a market substitute. It’s not even clear to me whether the republishing of stock imagery is pandemic, but many of the images posted on Pinterest are at least of similar quality to stock.
Flickr, one of the largest photography sites on the web, has already taken umbrage at Pinterest users’ unlicensed raiding of its members images, applying an “opt-out” code to those images account holders had set tighter sharing restrictions upon, and provided its account holders with a mechanism to prevent the use of their images on Pinterest. But Flickr, which is owned by Yahoo!, does not offer for sale any images of its account holders, who are largely hobbyists. Most Flickr account holders would thus have a difficult time showing market harm, unless they also proffered evidence that the images were generating revenue outside of Flickr. To the extent web traffic plays any part in the economic calculus, it is more likely that Pinterest would be found to have driven up traffic for Flickr’s account holders, and created for them more economic opportunities, as opposed to inflicting market harm.
Pinterest users also tend to post editorial photographs of the sort that regularly appear in fashion, craft, and home decor magazines. In many of these instances the question of ownership isn’t always clear–even if the photographer retained the rights, the magazine may enjoy a period of exclusivity including the right to sue. A lawsuit by rights-holding publications also seems unlikely, given how Pinterest drives traffic to the magazine and catalog sites, though such entities are better financially leveraged to engage in costly litigation than are photographers.
To the extent the photographer holds the rights or the rights revert back to him/her, the calculus switches–the photographer is less likely to benefit from the link back to the magazine or catalog, and consequently has a greater incentive to enforce the copyright, though individual photographers have fewer resources with which to litigate. Perhaps a trade association will emerge to fill this void; many exist but none on the scale of the RIAA or MPAA and none thus far who are visibly working to unite photography rights holders against Pinterest.
Despite Fagundes’s point that it essentially is the plaintiff’s burden to show market harm, defendants have thus far been able to show that their market improvement of an unlicensed work is favorably determinative.
This “market improvement” argument was primed to be fully tested after Google announced its extensive Library project in late 2004, in which Google would copy, republish, and either present or partially present more than one million books without first negotiating the rights of each one. Google would also provide a link to enable each books’ purchase and would remove the book from its library upon the complaint of any rights holder.
Not only are the public benefits to Google’s project clearly manifest, but the copyright holders also stood to benefit. [For a much deeper look into the copyright implications of the project, notably how it relates to the market harm, read Elisabeth Hanratty’s excellent student note here.] Google never explained, however, how the project could be squared with existing copyright law–Google was making full reproductions and displaying part, if not all, of the work without first even attempting to contact the rights holders. Even though it was a Win / Win / Win (Google / rights holders / public), it was also infringement.
Consequently, The Author’s Guild, a writers’ advocacy group, and then McGraw Hill, noted publisher, both sued Google, for infringement, seeking damages and injunctive relief. [Most relevant documents regarding the lawsuits can be found here]. Rather than test the uncertain waters of fair use, Google settled for approximately $125 million. [The complex, 323-page settlement [pdf here] established a fund from which authors whose works appear on Google would be compensated.]
Google’s decision to forego litigation is instructive here, because its Library model is analogous to Pinterest’s. Like Google, Pinterest is republishing works for which rights have not been negotiated, largely to the rights holders’ boon, and offering those rights holders who do not wish their works to be included, the ability to “opt out.” Unlike Google, however, Pinterest has not negotiated a nine-figure settlement with the rights holders. Pinterest faces continued uncertainty so long as its model persists and copyright law remains unchanged.
Despite my pessimism of Pinterest’s model with regard to how courts will view users’ infringements, I’m sympathetic to those users. Copyright law’s difficulty in adjusting to the sea changes brought on by the Internet is notorious, and each new tool that comes along seems to ratchet up the challenge. On Facebook, Internet users have been trained to become collectors and curators of various pieces of the internet–stories, videos, images–to reassemble those pieces on their “walls.” Pinterest leverages Facebook users’ intuitive training to a slightly different end; whereas Facebook users engage in discussions about this link and that video, which all serve to enhance a fair use argument, Pinterest users silently collect and curate, Pinning images of someone else’s butterfly, in the process stretching how we need to think about fair use.
My above analysis is far from exhaustive as to the copyright issues raised by Pinterest. I didn’t even discuss, for example, the huge issue of Pinterest’s potential liability as a secondary infringer–are they infringement inducers like Grokster? I’ll try to tackle that problem next. But I also didn’t factor in one very interesting realpolitik factor noted by reporter Anthony Wing Kosner in his Forbes Magazine article. Kosner points out that not only is Pinterest demographically skewed to women, it skews to slightly well-to-do, educated women living in the Midwest. [Kosner cites a demographic study by Ignite] In other words, the prototypical Pinterest user is the opposite of the less sympathetic most-commonly imagined Napster user. Both parties are infringers, but it’s harder to sell the idea of Pinterest users as thieves. Time will most likely tell because Pinterest’s popularity is increasing while copyright law stubbornly remains the same.