Right of Publicity

Keller v. Electronic Arts: The Ninth Circuit Adopts the Transformative-Use Test for College Athletes' Likenesses

On materially identical facts to Hart, the Ninth Circuit held that EA's photorealistic depiction of college football players in NCAA Football flunked California's transformative-use test and enjoyed no First Amendment shield.

February 18, 2025
A football stadium scoreboard glowing above an empty field at dusk
The court held that recreating an athlete in the very setting where he earned his fame is the opposite of transformation. Shutterstock
Educational content, not legal advice. This article explains general legal concepts. It does not create an attorney–client relationship. For your specific situation, consult a licensed attorney.

In re NCAA Student-Athlete Name & Likeness Licensing Litigation (Keller v. Electronic Arts Inc.), No. 10-15387 (9th Cir. July 31, 2013), is the Ninth Circuit’s foundational statement on how name, image, and likeness rights operate inside an interactive entertainment product. Samuel Keller, a starting quarterback for Arizona State and later Nebraska, sued Electronic Arts under California’s statutory and common-law rights of publicity for populating the NCAA Football franchise with avatars that mirrored real players down to jersey number, position, height, weight, home state, and playing style — while pointedly omitting their names. EA moved to strike the complaint under California’s anti-SLAPP statute, arguing that the games were protected speech. The U.S. District Court for the Northern District of California denied the motion, and a divided Ninth Circuit panel affirmed: Judge Bybee wrote for the majority, joined by Senior District Judge Quist sitting by designation, with Judge Sidney Thomas dissenting. The decision issued the same day as the panel’s companion ruling in Brown v. Electronic Arts, and two months after the Third Circuit’s parallel holding in Hart v. Electronic Arts.

At a glance

  • Case: In re NCAA Student-Athlete Name & Likeness Licensing Litigation (Keller v. Electronic Arts Inc.), No. 10-15387 (9th Cir. 2013)
  • Court / panel: U.S. Court of Appeals for the Ninth Circuit; Bybee, J. (majority), Quist, Sr. D.J. (by designation), Thomas, J. (dissenting)
  • Decided: July 31, 2013
  • Governing law: California statutory (Cal. Civ. Code § 3344) and common-law right of publicity
  • Posture: Appeal from denial of EA’s anti-SLAPP motion to strike under Cal. Code Civ. Proc. § 425.16
  • Holding: EA’s use of Keller’s likeness was not sufficiently transformative to claim First Amendment protection as a matter of law; denial of the motion to strike affirmed
  • Test applied: The transformative-use test from Comedy III Productions v. Gary Saderup, Inc., 25 Cal. 4th 387 (2001)

The anti-SLAPP procedure framed everything. Because EA invoked § 425.16, the plaintiff’s claim survived only if it arose from protected activity (conceded) and Keller could then show a probability of prevailing. EA’s First Amendment defenses were thus tested not as affirmative wins but as obstacles Keller had to clear. The majority’s task was to decide whether any recognized defense defeated the claim as a matter of law — and it concluded none did.

The transformative-use test, applied to NCAA Football

The doctrinal core of Keller is the panel’s selection and application of the transformative-use test. Drawing directly from the California Supreme Court’s decision in Comedy III, the court asked whether Keller’s likeness was “one of the raw materials from which an original work is synthesized,” or whether the depiction was “the very sum and substance of the work in question.” The inquiry borrows from copyright’s fair-use tradition but serves a different master: it balances the right of publicity against the First Amendment by measuring how much the defendant has added to, rather than merely appropriated from, the plaintiff’s identity.

Applied to NCAA Football, the test produced a clean loss for EA. The avatar bearing Keller’s attributes played his real-world position, for his real-world team, in a digitally rendered stadium, performing the same athletic feats through which he had earned public recognition. The court’s memorable formulation was that EA had recreated Keller “in the very setting in which he has achieved renown.” That context-preserving fidelity was the antithesis of transformation. The realism was not incidental to the product; it was the product’s commercial proposition.

EA pressed several theories the panel rejected in turn. It argued that user-customization features — the ability to alter an avatar’s appearance and attributes — supplied the necessary creative transformation. The court refused to credit hypothetical edits, reasoning that the relevant depiction was the one EA shipped, in which Keller’s likeness appeared in recognizable form. EA also urged the court to look at the game holistically, pointing to the creative graphics, virtual venues, music, and interactive mechanics surrounding the avatars. The majority declined: transformation must inhere in the depiction of the plaintiff, not in the expressive architecture wrapped around an otherwise literal reproduction. A creative shell does not launder an appropriated core.

Why the court rejected EA’s competing defenses

Beyond transformative use, EA invoked a cluster of First Amendment defenses the Ninth Circuit had developed in adjacent contexts, and the panel dispatched each. EA argued for the Rogers v. Grimaldi artistic-relevance test imported from trademark law. The majority held Rogers inapplicable to a right-of-publicity claim, reasoning that a test designed to gauge consumer confusion about the source of an expressive work does not measure the distinct harm at the heart of publicity — the misappropriation of identity itself. (The companion Brown decision, by contrast, applied Rogers precisely because Jim Brown’s claim sounded in the Lanham Act, illustrating that the choice of test tracks the cause of action, not the medium.)

EA next invoked the “public interest” and “public affairs” defenses, contending that depicting real athletes reported on matters of public concern. The panel found these inapt: those defenses protect the reporting of information, not the wholesale commercial reproduction of a person’s likeness in a game. Finally, EA raised California’s statutory exemption and a transformative “incidental use” theory; the court held the use was central, not incidental, because the realistic depiction of identifiable athletes was the very draw of the franchise.

Judge Thomas dissented. Accepting the transformative-use framework, he would have examined the game holistically and concluded that its “transformative and creative elements” — the interactive design, the vast roster of customizable players, the simulated seasons — predominated over any literal, imitative depiction of a single athlete. His disagreement was not about the test but about the lens: whether transformation is assessed at the level of the discrete likeness or the work as a whole. That narrow-versus-wide-lens dispute, also surfaced in the Hart dissent, remains the most consequential fault line in the doctrine.

A circuit convergence with industry consequences

Keller did not stand alone. The same week, the Third Circuit had decided Hart v. Electronic Arts on materially identical facts, adopting the same transformative-use test and reaching the same result. Two federal courts of appeals, applying the law of two different states to the same franchise, agreed that photorealistic, unlicensed depiction of identifiable athletes was not protected expression as a matter of law. That convergence carried outsized practical weight. EA settled the consolidated student-athlete litigation, and the NCAA Football franchise went dark for years before its eventual relaunch under a paid NIL-licensing model. The doctrine, in other words, reordered a licensing market: once realism could not be defended as transformation, the only path back to the product was consent and compensation.

Open questions

  • Is transformation measured by the discrete depiction of the plaintiff or by the expressive work as a whole? Keller’s majority chose the former; Judge Thomas’s dissent and several commentators favor the latter, and the answer frequently decides cases.
  • How does the transformative-use test square with Zacchini v. Scripps-Howard Broadcasting Co., 433 U.S. 562 (1977), the only Supreme Court right-of-publicity decision, which permitted a claim despite newsworthy content?
  • After Jack Daniel’s Properties v. VIP Products, 599 U.S. 140 (2023), narrowed Rogers in the trademark setting, is there any remaining reason to prefer Rogers over transformative use for likeness claims even where the two could overlap?
  • Does generative AI, which can synthesize a recognizable-yet-novel likeness, fit the transformative-use rubric at all, or does it demand new analytic vocabulary?

Implications

  • Realism is a liability, not a shield. A product that sells on the fidelity of its depiction of real people has the weakest possible transformative-use defense; the very feature that drives revenue defeats the constitutional argument.
  • Customization does not transform. Courts evaluate the as-shipped depiction, not a user’s hypothetical ability to alter it.
  • The cause of action selects the test. Keller applied transformative use to a publicity claim while its companion Brown applied Rogers to a Lanham Act claim — practitioners must match framework to claim, not to medium.
  • Anti-SLAPP raises the stakes early. Because EA’s defenses were tested at the motion-to-strike stage, the decision shows that publicity claims against expressive works can survive threshold dismissal where the depiction is literal.
  • Doctrine can move markets. The Keller/Hart convergence drove settlement and a licensing overhaul, foreshadowing the consent-based NIL economy that now governs college athletics.

Frequently asked questions

How is Keller different from Hart? They are companion cases on nearly identical facts decided within two months of each other. Hart arose in the Third Circuit under New Jersey law; Keller arose in the Ninth Circuit under California law and reached the court through an anti-SLAPP motion. Both adopted the Comedy III transformative-use test and both held that EA’s NCAA Football depictions were not transformative. Their agreement is what gave the rule its national force.

Why didn’t EA win under Rogers v. Grimaldi? The Keller majority held that Rogers, a trademark test focused on consumer confusion about a work’s source, does not fit a right-of-publicity claim, whose gravamen is the appropriation of identity rather than deception about endorsement. The same panel did apply Rogers the same day in Brown — because Jim Brown sued under the Lanham Act, where confusion is the operative harm.

Did the names matter, given that EA omitted them? No. The court treated likeness as a composite of recognizable attributes — number, position, physique, team, home state, statistics — not as a function of the player’s name. Stripping the name while preserving the recognizable whole did not defeat the claim.

Authorities and sources