When the Avatar Is You: Lohan v. Take-Two and the Recognizability Line

New York's highest court held that a video-game avatar can be a 'portrait' under the right of publicity, but only if the public can actually recognize the plaintiff in it.

In Lohan v. Take-Two Interactive Software, Inc., 2018 NY Slip Op 02208 (N.Y. Ct. App. Mar. 29, 2018), the New York Court of Appeals confronted a question that had quietly been building under decades of likeness law: can a digitally rendered character — an avatar — be a person’s “portrait”? Lindsay Lohan said the character “Lacey Jonas” in Rockstar Games’ Grand Theft Auto V was a thinly veiled version of her, complete with a beach setting, blonde hair, and a “signature peace-sign pose.” Writing for a unanimous court, Judge Eugene Fahey gave defendants a loss on the doctrine and a win on the facts: yes, an avatar can be a statutory “portrait,” but no, this one was not recognizable as Lohan. The complaint was dismissed.

At a glance

Case: Lohan v. Take-Two Interactive Software, Inc., 2018 NY Slip Op 02208; 31 N.Y.3d 111 (Mar. 29, 2018) (Fahey, J.).
Court: New York Court of Appeals (the state’s highest court).
Issue: Whether a graphical avatar in a video game can be a “portrait” under N.Y. Civil Rights Law §§ 50–51, and whether the GTA V character was recognizable as Lindsay Lohan.
Holding: A computer-generated image can constitute a “portrait” within the statute — but the disputed character was “not recognizable as plaintiff” and was a generic, satirical depiction, so the claim failed as a matter of law.
Why it matters: It modernized New York’s narrow statutory likeness right for digital media while erecting recognizability as the decisive gatekeeper for avatar-based claims.

New York’s statute, not a common-law tort

A first principle frames everything in Lohan: New York does not recognize a freestanding common-law right of publicity. It has a statute — Civil Rights Law §§ 50 and 51 — that prohibits the use of a living person’s “name, portrait, picture or voice” for “advertising purposes or for the purposes of trade” without written consent. The provision was enacted in 1903 in reaction to Roberson v. Rochester Folding Box Co., where the Court of Appeals had refused to recognize a privacy interest in a woman whose face was used on flour advertisements. The statute is therefore remedial but also deliberately bounded: it reaches enumerated attributes used for commerce, not every evocation of a public figure.

That statutory framing matters for influencers and entertainers because it sets the terms of the fight. The question is never “did this remind people of me?” It is “did the defendant use my name, portrait, picture, or voice?” In a video game, the contested attribute is almost always a “portrait” or “picture,” which forced the court to decide whether a rendered avatar qualifies at all.

Avatars as “portraits”: the doctrinal expansion

The court answered yes. A “portrait,” it explained, is not limited to photographs; it has long encompassed artistic “representations” of a person, and “[t]hat an image is a portrait does not depend on the medium.” A digital avatar is just a representation rendered by software rather than by a brush or a camera. Refusing to extend the statute to game characters would let the most pervasive medium of modern persona — interactive entertainment — escape a likeness regime built for an analog world.

This is the part of Lohan with the longest reach. By holding that the “portrait” category is medium-neutral, the court future-proofed §51 against the argument that a character is “merely” code. The same logic now travels to motion-captured athletes, scanned actors, and — increasingly — AI-generated faces. If the output is recognizable as a particular person and is used in trade, the rendering technology is irrelevant.

Recognizability: the gate that closed

Having opened the door, the court shut it on Lohan. The decisive question was whether the Lacey Jonas avatar was “recognizable as plaintiff.” It was not. The character, the court found, was “not recognizable as Lindsay Lohan” but rather “a generic artistic depiction of a ‘twenty something’ woman without any particular identifying physical characteristics.” The images were “indistinct, satirical representations … of the lifestyle of young women,” a form of “cultural comment” rather than an appropriation of an identifiable individual.

Two analytical moves are bundled here. The first is the recognizability requirement itself: a §51 portrait claim requires that the plaintiff be identifiable from the image, judged by the perception of the audience, not by the plaintiff’s subjective conviction or by the defendant’s private intent. The second is a content characterization: where the depiction is a generalized, satirical archetype rather than a specific likeness, it both fails recognizability and edges into expression that the First Amendment and the statute’s “trade” limits protect. A game is not an “advertisement,” and the use of an evocative type — the entitled influencer, the party girl — is precisely the kind of cultural commentary the statute was never meant to police.

Importantly, the court did not rest on a transformative-use balancing test of the sort California courts apply. It resolved the case on the threshold statutory element. That makes Lohan a cleaner, more predictable rule for New York defendants: win on recognizability and you never reach the harder First Amendment weighing.

Open questions

How specific must the borrowing be? Lohan pointed to hair, a beach scene, and a pose, yet lost on recognizability. The opinion does not say how many specific, individualized markers — a tattoo, a catchphrase, a unique silhouette — would flip the result.
Who decides recognizability, and when? The court treated it as resolvable on the pleadings here, but a closer avatar might present a jury question, raising the risk that survivable claims turn on early, fact-intensive characterizations of an image.
Does the satire framing survive precise replicas? The court leaned on the “generic” and “satirical” nature of the character. A scanned, photorealistic double used for trade would not be generic — would the same expressive breathing room apply?

Implications

For game studios: Building characters from archetypes rather than identifiable individuals is a defense, not just a design choice. Generic depictions of a “type” are far safer than renderings keyed to one real person’s distinctive features.
For influencers and entertainers: Feeling personally referenced is not enough. A New York §51 claim requires that the public recognize you in the image, and resemblance to a lifestyle or vibe will not satisfy that.
For AI and likeness pipelines: Because “portrait” is medium-neutral, AI-generated faces and motion-captured doubles fall squarely within §51 when they are recognizable and used in trade — the technology offers no immunity.
For litigants generally: Recognizability is a front-loaded, potentially dispositive element in New York. Plaintiffs should plead concrete identifying features; defendants should attack identifiability before reaching the First Amendment.

Frequently asked questions

Did Lohan lose because video-game characters can never be portraits? No — the opposite. The court held that an avatar can be a “portrait” under §§ 50–51. Lohan lost because the specific character was not recognizable as her and was a generic, satirical depiction.

Does this decision recognize a broad right of publicity in New York? No. New York has no common-law right of publicity; it has only the statutory privacy/publicity right in Civil Rights Law §§ 50–51, which reaches enumerated attributes used for advertising or trade. Lohan applies that narrow statute, it does not expand it into a general identity right.

Could a more detailed character have won? Possibly. The court stressed that the avatar lacked “particular identifying physical characteristics.” A depiction tied to a plaintiff’s specific, recognizable features — used in trade and not as generalized satire — could clear the recognizability threshold the court applied.