Video game is transformative use of player's image

Hart v. Electronic Arts, Inc., --- F. Supp. 2d ---, 2011 WL 4005350 (D.N.J.)

In a long and sure to be much discussed opinion, the district court handed EA a significant victory as a First Amendment speaker, despite using a balancing test to determine EA’s interests that has no rightful place in a regulation of noncommercial speech. (Correspondingly, the court refused to hold that the right of publicity is content based, even though it plainly is by the only official definitions of “content” that distinguish “content” from “viewpoint.” Here I have less complaint, because I don’t think that “content” is a useful category for determining the level of scrutiny to be applied to a speech regulation in the way that “viewpoint” is, as I will detail below.)

Hart, a former college football player at Rutgers, sued EA for misappropriating his likeness and identity on behalf of a class of those similarly situated.

EA produces a video game series, NCAA Football. Each game permits users to “manipulate the actions of over 100 college football teams and thousands of virtual players in a virtual world with simulated games.” The teams are identifiable by name, and players are identified by jersey number and position. Users can edit game data and given players surnames, which appear on their jerseys. “Each virtual player's unique attributes, including personal characteristics (height, weight, athletic ability), accessories (helmet visor, wristband), physical abilities (speed and agility, throwing arm, passing accuracy), and biographical details (place of origin) can also be edited by the user.” (I’m confused, because a footnote also says that the player’s name, position, number, and handedness can be changed, but not the player’s home state, hometown, team, and year.) Users can also download custom rosters created by other players, and one place they can do that is Teambuilder on EA’s website. “Some rosters available on these websites seek to replicate actual current and former football team rosters.” School names, team names, uniforms, logos, and stadium fight songs “appear in the game through licensing agreements between EA and the NCAA's licensing agent.”

Users control the players, and they can play against the game or against other players. In Dynasty mode, a user controls a college program for up to thirty seasons. Players in Dynasty mode have the "year-round responsibilities of a college coach, such as recruiting virtual high school players out of a random-generated pool of athletes."

Hart claimed that the games depicted a "virtual" player that had been designed to replicate his physical attributes (height and weight), as well as his football skills, and that EA had used video footage of him playing in a Rutgers University Football game to promote the game. The virtual player at issue wore Hart’s jersey number and his speed and agility rating, passing accuracy rating, and arm strength all allegedly reflected actual footage of Hart in his 2005 season. He argued that the games allowed users to “simulate the college football playing experience by stepping into the shoes of Rutgers' QB Ryan Hart, and other college football players, where fans can mimic Plaintiff's style and movements and play against Plaintiff's actual opponents.” EA didn’t dispute that a photo of Hart throwing a pass appeared in a photo montage inside NCAA Football 2009, which is only seen when a user selects Rutgers as his or her favorite team.

Hart alleged that EA misappropriated his identity with the intent of increasing sales, because heightened realism in video games translates directly into increased revenue. Further, he alleged that EA wrongfully failed to prevent users from uploading and downloading team rosters with real players’ names, again to heighten the authenticity and realism of a “true NCAA football experience.”

Although the court initially refused to dismiss the right of publicity claims under New Jersey law, it now granted summary judgment (converting EA’s motion to dismiss).

Brown v. Entertainment Merchants Ass'n, 131 S.Ct. 2729 (2011), confirmed that video games are entitled to First Amendment protection. Hart argued that Brown was irrelevant because it involved a content-based statute. EA rejoined that the right of publicity is content-based as well, and that Brown confirms that video games are not commercial speech. The court agreed with EA’s latter proposition but not the former. Applying Facenda, Hart “cannot reasonably contend that the NCAA Football games constitute commercial speech.” Facenda involved an infomercial for Madden Football touting the soon-to-be released game. The video in that case promoted another creative work, the video game. “Here, by contrast, the speech is the video game that is being sold. It is not a separate instance of speech that promotes the purchase of another work.” Hart provided no evidence that his image was used in ads for the game (as opposed to being part of the game).

The court quoted the Third Circuit’s statement that a content-based speech restriction is one that regulates speech “when the specific motivating ideology or the opinion or perspective of the speaker is the rationale for the restriction.” Galena v. Leone, 638 F.3d 186, 199 (3d Cir.2011). Comment: um, no. Those are viewpoint-based restrictions, as the references to “ideology,” “opinion,” and “perspective” as opposed to “content” or “subject matter” indicate. Content-based restrictions can be driven by viewpoint bias—see, e.g., the execrable “No Promo Homo” proposals floating around—but then we should care about the viewpoint bias. The right of publicity, unlike a regulation of the volume of loudspeakers or a ban on demonstrations of any kind at the public zoo (content-neutral), requires an examination of the particular content at issue to figure out if a violation has occurred. The problem is that “content-based” is too broad a spectrum to be meaningful; some content-based regulations (e.g., regulation of false and misleading commercial speech) are perfectly fine and some aren’t. The breadth of the covered topic should make a difference to the analysis, and the right of publicity may be closer to copyright than to mentions of homosexuality. But at best this is semi-content-neutral regulation.

And indeed, citing Thomas F. Cotter, et al., Integrating the Right of Publicity with First Amendment and Copyright Preemption Analysis, 33 Colum. J. of Law & the Arts 165 (Winter 2011), the court pointed out that most courts haven’t analyzed the right of publicity as content-based. Instead, courts apply one of several balancing tests “unique to intellectual property-related cases” to determine whether the First Amendment limits a right of publicity claim, rather than applying strict or intermediate scrutiny. And since we’ve always given IP—even IP with limited historical foundation and massively expanding scope—special exemption from First Amendment scrutiny, the court continued to do so, formulating its job as to choose between existing tests. Or actually, not, because under either Rogers or transformativeness the result was the same here.

Transformativeness is borrowed from copyright’s fair use doctrine, which helps reconcile copyright and the First Amendment. Citing David Tan, Political Recoding of the Contemporary Celebrity and the First Amendment, 2 Harv. J. Sports & Ent. L. 1, 25-26 (2011), the court noted that transformativeness has been criticized as lacking objective guidelines, specifically pointing to the role the fame of the artist might (wrongly) play. As Tan puts it: “the cryptic judicial comments that literal depictions like Andy Warhol's silkscreens of celebrities may also be transformative if they carry a particular social message lend little guidance to how a court may meaningfully determine what constitutes the criteria for transformative use.” Tan continued, “the test is focused on visual transformation which can be overprotective of art and entertainment that contribute little to the discussion of public issues, but underprotective of political speech which may be contextually transformative (because of its recoding) though not visually transformative.” Copyright’s fair use doctrine is pretty nebulous, and borrowing transformativeness from it is courting danger.

But, the court rejoined, fair use “nonetheless remains the statutory law for copyright matters and the U.S. Supreme Court found it appropriate in Zacchini to analogize the right of publicity to federal copyright law.” Plus, Eldred says fair use helps make copyright constitutional, so even if transformativeness forces judges to be art critics, “a degree of artistic interpretation” is okay. “[T]he test's rooting in the fair use doctrine and its consideration of the extent to which an image is copied fairly takes into account the competing right of publicity and First Amendment interests.”

The court reviewed various California cases, and found two particularly instructive: Kirby v. Sega of America, Inc., 144 Cal. App. 4th 47 (Cal. App. 2006), holding that a video game character fashioned after the plaintiff, a celebrity singer, was a transformative use. Kirby was known for the phrase "ooh la la." The video game character was named "Ulala," and shared some physical similarities, clothing style, and catchphrases with the singer. However, Ulala had a distinct, anime-derived physique, and often varied hair and clothing. The character was also a 25th-century reporter, and didn’t dance the way Kirby did. So the character was overall transformative.

By contrast, No Doubt v. Activision, Inc., 192 Cal. App. 4th 1018 (Cal. App. 2011), dealt with the avatars who were literal recreations of the band members, who had posed for photography for the video game developer. These were not transformative, but instead “immutable images of the real celebrity musicians.”

The court found this case a closer call. Hart’s NCAA Football virtual player plays college football, just like Hart did. But he’s not an exact replica that can’t be altered by the user. Rather, he can be altered in many ways, “from his personal characteristics (height, weight, athletic ability), to his accessories (helmet visor, wristband),” along with his “physical abilities (speed and agility, throwing arm, passing accuracy), attributes, and certain biographical details (right handed/left handed).”

Still, “[p]lacing present and former college athletes, including Hart, into the fittingly-titled NCAA Football game setting strongly suggests that the goal of the game is to capitalize upon the fame of those players.” But heightened realism doesn’t mean that the use isn’t transformative, since transformativeness can result from added creative elements that operate on the celebrity depiction. “Nevertheless, a game developer that bases its work on real players, in the context of the games that bring them notoriety, may walk a fine line between using reality as a building block for the developer's own creative work and exploiting the hard-earned reputations of college players for its own profit.” (Okay, explain to me how the court can say that video games are noncommercial speech, like movies and books, and then also say this without putting biography and the like at risk. I know the answer—writing is real creative work, unlike making pictures, so it will always count as transformative—but it’s a bad answer.)

The court was sensitive to the plight of college players “who are prohibited by NCAA bylaws from entering into licensing agreements and other ‘commercial opportunities’ during their playing years,” and noted Hart’s argument that EA pays pro football players for the use of their images as well as colleges for the use of their marks, uniforms, and logos. But unfairness didn’t resolve the First Amendment question.

“Viewed as a whole, there are sufficient elements of EA's own expression found in the game that justify the conclusion that its use of Hart's image is transformative and, therefore, entitled to First Amendment protection.” The game has several creative elements apart from Hart’s image: "virtual stadiums, athletes, coaches, fans, sound effects, music, and commentary, all of which are created or compiled by the games' designers." Even Hart’s virtual image alone was transformative because users could change its features, statistics, and teammates. Moreover, the court found the C.B.C. case persuasive that the use of player names, statistics, and biographical data is entitled to First Amendment protection because the data are in the public domain. Facts aren’t entitled to copyright protection, and are available to everyone to use; this is another IP rule that protects First Amendment interests.

The court specified that it wasn’t the user’s alteration of Hart’s image that controlled, but that EA created the mechanism by which the virtual player could be altered, as well as offering multiple options for each image: height, weight, hairstyle, face shape, body size, muscle size, and complexion. There were eight possible hairstyles, 24 face options, 16 face shape options, and so on. This made the game a transformative use.

But, you might say, using Hart’s unaltered image as the starting point isn’t transformative. “The problem with this argument is that it fails to fully take into account the distinctive interactive nature of video games…. [A] user's interaction with a video game is one of the means by which video games communicate ideas and social messages.” What about the photo of Hart throwing a pass? That’s a fleeting component of the video game as a whole, and, “[b]ecause the photograph comprises such a small portion of the entire work, it cannot be said that the work itself fails to incorporate transformative elements.”

Hart attached screenshots along with photos of him taken during his college career and argued that they showed him and the virtual player in the same positions. The court disagreed that the positions were the same, but that wasn’t the real point. “That a few still screenshots in the context of a larger video game reflect his image does not undercut the existence of the additional creative elements of the game, discussed above, and the variable permutations of his image, created by EA's artists, for use in the game.”

Hart argued that EA deliberately set out to create "the most realistic authentic [football] performance [so that the user may] feel what it's like to cover the field and play at the most elite level in college football," in EA’s own words. But realism isn’t the focus, transformativeness is. (Implicitly, then, realism can be transformative?)

Keller, a N.D. Cal. case that is currently on appeal, found to the contrary. The court here distinguished Keller because it was decided on a motion to dismiss and didn’t consider the declarations describing the games. Also, Keller failed to address the interactivity of the virtual player using variations created by EA artists. The court found this significant “because it suggests that the goal of the game is not for the user to ‘be’ the player, as in No Doubt, where the virtual player could not be altered. The malleability of the player's image in NCAA Football suggests, instead, that the image serves as an art-imitating-life starting point for the game playing experience.” Thus, the image serves as one of the “raw materials” for the new work, and the celebrity depiction isn’t the “sum and substance” of the work (quoting Comedy III). “Moreover, one could argue that even the technology that permits users to alter a player's image is itself a noteworthy, expressive attribute of the game.”

In addition, the court here disagreed with Keller’s focus solely on the challenged image as opposed to the work as a whole.

The court then turned to the Rogers test that protects artistically relevant uses in expressive works, which approach the court found inferior to transformativeness. First, the court thought that Rogers didn’t explain the genesis of the test, though it may have been borrowed from NY’s “real relationship” test used to figure out when newsworthiness trumps publicity rights: there’s no violation of the right of publicity unless the image has no “real relationship” to a news item or unless the item is an ad in disguise. This test protects uses of an image that are noncommercial or not used for purposes of trade, but it’s been criticized as manipulable depending on the level of generality (that is, courts may differ on whether there’s a sufficient connection between a newsworthy topic and this particular person). Yet Rogers expands on the “real relationship” test and doesn’t even require newsworthiness, which the court found perplexing. (The underlying problem here is that there shouldn’t be a “real relationship” test, only a noncommerciality/purposes of trade test; once we’ve added the further requirement of relevance, though, and once we admit that art gets First Amendment protection just like news does, artistic relevance makes sense as the equivalent of newsworthiness.)

The court here thought that Rogers imported this test without explaining why it was helpful in the Lanham Act context, and that subsequent courts had jogged back and forth in applying Rogers to right of publicity claims as well as Lanham Act claims without sufficient justification. In fact, Rogers itself pointed out that the right of publicity has no confusion requirement, making it potentially more expansive than the Lanham Act (and thus potentially more threatening to free speech). The court concluded that there were actually two Rogers tests, one for each cause of action, and questioned the wisdom of applying a trademark-based test to the right of publicity without accounting for the latter’s absence of a confusion requirement. (Well, sure, but I don’t see how that counsels picking transformativeness over artistic relevance; if anything, it counsels protecting any use that is either transformative or artistically relevant against a right of publicity claim, at least for noncommercial speech.)

One arguable benefit of Rogers, the court continued, was that it was more straightforward and required less artistic interpretation, but it was still subject to manipulation. (Surprisingly, the court didn’t identify an instance of such manipulation, though Parks is the clear winner in that regard.) Still, three other circuits have used Rogers, and a couple of district court opinions have applied it specifically to video games.

So, assuming for the sake of argument that Rogers supplies the proper standard, EA still wins. Using Hart’s image “has great relevance to the game itself, which is set on a college football field and revolves around the playing of virtual football.” And the use is not misleading; Hart didn’t even try to claim false endorsement. Nor was the use here a disguised commercial advertisement. “[T]he use of his image is part of an expressive act by EA that might draw upon public familiarity with Hart's college football career but does not explicitly state that he endorses or contributes to the creation of the game.” Use of Hart’s identity in an expressive work relevant to that identity is legitimate. (Yes. Yes, it is, which is why the transformativeness test is underinclusive in this type of situation.)

So, the court concluded, “the transformative test best encapsulates the type of nuanced analysis required to properly balance the competing right of publicity and First Amendment interest.” (Unsurprisingly, I don’t get this in the slightest, and it seems to me that the court’s own comments about First Amendment protection for free dissemination of publicly known facts necessarily entail that transformativeness is not sufficient to protect First Amendment interests in noncommercial speech.) But EA wins no matter what so the court didn’t really need to decide the issue.

Concluding thoughts: in many ways, this is a model of a particular type of opinion: it is incredibly careful, attempting to resolve questions that have divided courts on practically identical facts while breaking no new ground whatsoever. The defects of its approach stem from the inherent flaws of the modern right of publicity, which claims property rights in others’ expression. If one purpose of art is to hold a mirror up to nature, the celebrity claims to own the mirror by virtue of being reflected in it.