OPINION
Plaintiff — Appellant James “Jim” Brown alleges that Defendant — Appellee Electronic Arts, Inc. (“EA”) has violated § 43(a) of the Lanham Act, 15 U.S.C. *1239§ 1125(a), through the use of Brown’s likeness in EA’s Madden NFL series of football video games. In relevant part, § 43(a) provides for a civil cause of action against:
[a]ny person who, on or in connection with any goods or services, or any container for goods, uses in commerce any word, term, name, symbol, or device, or any combination thereof, or any false designation of origin, false or misleading description of fact, or false or misleading representation of fact, which ... is likely to cause confusion, or to cause mistake, or to deceive as to the affiliation, connection, or association of such person with another person, or as to the origin, sponsorship, or approval of his or her goods, services, or commercial activities by another person[.]
15 U.S.C. § 1125(a)(1). Although claims under § 43(a) generally relate to the use of trademarks or trade dress to cause consumer confusion over affiliation or endorsement, we have held that claims can also be brought under § 43(a) relating to the use of a public figure’s persona, likeness, or other uniquely distinguishing characteristic to cause such confusion.1
Section 43(a) protects the public’s interest in being free from consumer confusion about affiliations and endorsements, but this protection is limited by the First Amendment, particularly if the product involved is an expressive work. Recognizing the need to balance the public’s First Amendment interest in free ex•pression against the public’s interest in being free from consumer confusion about affiliation and endorsement, the Second Circuit created the “Rogers test” in Rogers v. Grimaldi, 875 F.2d 994 (2d Cir.1989). Under the Rogers test, § 43(a) will not be applied to expressive works “unless the [use of the trademark or other identifying material] has no artistic relevance to the underlying work whatsoever, or, if it has some artistic relevance, unless the [use of trademark or other identifying material] explicitly misleads as to the source or the content of the work.” Id. at 999. We adopted the Rogers test in Mattel, Inc. v. MCA Records, Inc., 296 F.3d 894 (9th Cir.2002).
Applying the Rogers test, the district court in this case granted EA’s motion to dismiss Brown’s Lanham Act claim, finding that Brown had not alleged facts that satisfied either condition that allow a § 43(a) claim to succeed under the Rogers test. Brown v. Elec. Arts, Inc., No. 2:09-cv-01598, 2009 WL 8763151, at *3-5, 2009 U.S. Dist. LEXIS 131387, at *8-15 (C.D.Cal. Sept. 23, 2009). Brown appealed, challenging the applicability of the Rogers test, the district court’s analysis under the Rogers test, and the suitability of his case for resolution without additional factfinding. We affirm the district court’s decision.
I
Jim Brown is widely regarded as one of the best professional football players of all *1240time. He starred for the Cleveland-Browns from 1957 to 1965 and was inducted into the National Football League (“NFL”) Hall of Fame after his retirement. After his NFL career, Brown also achieved success as an entertainer and public servant. There is no question that he is a public figure whose persona can be deployed for economic benefit.
EA is a manufacturer, distributor and seller of video games and has produced the Madden NFL series of football video games since 1989. The Madden NFL series allows users of the games to control avatars representing professional football players as those avatars participate in simulated NFL games. In addition to these simulated games, Madden NFL also enables users to participate in other aspects of a simulated NFL by, for example, creating and managing a franchise. Each version of Madden NFL includes the current year’s NFL teams with the teams’ current rosters. Each avatar on a current team is designed to mirror a real current NFL player, including the player’s name, jersey number, physical attributes, and physical skills. Some versions of the game also include historical and all-time teams. Unlike for players on the current NFL teams, no names are used for the players on the historical and all-time teams, but these players are recognizable due to the accuracy of their team affiliations, playing positions, ages, heights, weights, ability levels, and other attributes. Although EA enters into licensing agreements with the NFL and NFL Players Association (“NFLPA”) for its use of the names and likenesses of current NFL players, Brown, as a former player, is not covered by those agreements and has never entered into any other agreement allowing EA to use his likeness in Madden NFL. Brown asserts that EA has used his likeness in several versions of the game dating back at least to 2001 but that he has never been compensated.
Brown brought suit in the United States District Court for the Central District of California, claiming that EA’s use of his likeness in the Madden NFL games violated § 43(a) of the Lanham Act. Brown also brought claims under California law for invasion of privacy and unfair and unlawful business practices. EA filed a motion to dismiss pursuant to Rule 12(b)(6) of the Federal Rules of Civil Procedure, and the district court applied the Rogers test and dismissed Brown’s Lanham Act claim. Brown, 2009 WL 8763151, at *3-5, 2009 U.S. Dist. LEXIS 131387, at *9-15. The district court declined to exercise supplemental jurisdiction over the state-law claims. Id. at *5-6, 2009 U.S. Dist. LEXIS 131387, at *15-16. Brown filed a timely appeal of the dismissal of his Lanham Act claim.2 We have jurisdiction pursuant to 28 U.S.C. § 1291. We review the district court’s dismissal de novo. Kahle v. Gonzales, 487 F.3d 697, 699 (9th Cir.2007).
II
The legal issues raised by this case are not novel, but their lack of novelty should *1241not be mistaken for lack of difficulty. Significant judicial resources, including the resources of this court, have been expended trying to find the appropriate balance between trademark and similar rights, on the one hand, and First Amendment rights, on the other. Brown suggests that the case law has produced a lack of clarity as to the appropriate legal framework to apply in this case and urges us to consider the “likelihood of confusion” test and the “alternative means” test in addition to the Rogers test. We are convinced that the Rogers test remains the appropriate framework.
A decade ago, in Mattel, Inc. v. MCA Records, Inc., we adopted the Rogers test as our method for balancing the trademark and similar rights protected by § 43(a) of the Lanham Act against First Amendment rights in cases involving expressive works. MCA, 296 F.3d at 902. Although MCA concerned the use of a trademark in the title of an expressive work, and the language of the MCA opinion did not make it clear that we were adopting the Rogers test for cases where the trademark or other identifying material in question was used in the body of a work rather than in the title, we clarified in E.S.S. Entertainment 2000, Inc. v. Rock Star Videos, Inc. that application of the Rogers test was not dependent on the identifying material appearing in the title but “also applfies] to the use of a trademark in the body of the work.” 547 F.3d 1095,1099 (9th Cir.2008). We have consistently employed the Rogers test in § 43(a) cases involving expressive works since MCA, including where the trademark or other identifying material in question was used in the body of a work rather than in the title. See, e.g., id.; Mattel, Inc. v. Walking Mountain Prods., 353 F.3d 792 (9th Cir.2003).
The Rogers test is reserved for expressive works. Even if Madden NFL is not the expressive equal of Anna Karenina or Citizen Kane, the Supreme Court has answered with an emphatic “yes” when faced with the question of whether video games deserve the same protection as more traditional forms of expression. In Broum v. Entertainment Merchants Ass’n, the Court said that “[l]ike the protected books, plays, and movies that preceded them, video games communicate ideas— and even social messages — through many familiar literary devices (such as characters, dialogue, plot, and music) and through features distinctive to the medium (such as the player’s interaction with the virtual world)” and that these similarities to other expressive mediums “suffice[ ] to confer First Amendment protection.” — U.S. -, 131 S.Ct. 2729, 2733, 180 L.Ed.2d 708 (2011). Although there may be some work referred to as a “video game” (or referred to as a “book,” “play,” or “movie” for that matter) that does not contain enough of the elements contemplated by the Supreme Court to warrant First Amendment protection as an expressive work, no version of Madden NFL is such a work. Every version of the game features characters (players), dialogue (between announcers), plot (both within a particular simulated game and more broadly), and music. Interaction between the virtual world of the game and individuals playing the game is prevalent. Even if there is a line to be drawn between expressive video games and non-expressive video games, and even if courts should at some point be drawing that line, we have no need to draw that line here.3 Each version of Madden NFL is an expressive work, and our precedents dictate that we apply the Rogers test in § 43(a) cases in*1242volving expressive works. Brown acknowledges that Rogers may apply here, but he argues that the “likelihood of confusion” test, exemplified by Dr. Seuss Enterprises, L.P. v. Penguin Books USA, Inc., 109 F.3d 1394 (9th Cir.1997), or the “alternative means” test, exemplified by International Olympic Committee v. San Francisco Arts & Athletics, 781 F.2d 733 (9th Cir.1986), reh’g en banc denied, 789 F.2d 1319 (9th Cir.1986), aff'd on other grounds, S.F. Arts & Athletics, Inc. v. U.S. Olympic Comm., 483 U.S. 522, 107 S.Ct. 2971, 97 L.Ed.2d 427 (1987), are also relevant. We disagree. We have previously rejected the “likelihood of confusion” test as “failing] to account for the full weight of the public’s interest in free expression” when expressive works are involved. MCA, 296 F.3d at 900. The “alternative means” test was rejected for the same reason in Rogers itself, 875 F.2d at 999, a position we approved by adopting the Rogers test in MCA The only relevant legal framework for balancing the public’s right to be free from consumer confusion about Brown’s affiliation with Madden NFL and EA’s First Amendment rights in the context of Brown’s § 43(a) claim is the Rogers test.
Ill
Rogers involved a suit brought by the famous performer Ginger Rogers against the producers and distributors of Ginger and Fred, a movie about two fictional Italian cabaret performers who imitated Rogers and her frequent performing partner Fred Astaire. Rogers, 875 F.2d-at 996-97. Among Rogers’ claims was that the use of her name in the title of the movie violated § 43(a) by creating' the false impression that she was involved with the film. Id. at 997. Recognizing that enforcing § 43(a) in this context might constrain free expression in violation of the First Amendment, the Second Circuit asserted that the Lanham Act should be “applied] to artistic works only where the public interest in avoiding consumer confusion outweighs the public interest in free expression.” Id. at 999. The Rogers court introduced a two-pronged test, under which the Lanham Act should not be applied to expressive works “unless the [use of the trademark or other identifying material] has no artistic relevance to the underlying work whatsoever, or, if it has some artistic relevance, unless the [trademark or other identifying material] explicitly misleads as to the source or the content of the work.” Id. ‘
A
As we explained in E.S.S., a case with similar facts to Brown’s case in which we applied the Rogers test to a § 43(a) *1243claim related to the use of the likeness of a Los Angeles strip club in the video game Grand Theft Auto: San Andreas, “the level of [artistic] relevance [of the trademark or other identifying material to the work] merely must be above zero” for the trademark or other identifying material to be deemed artistically relevant. 547 F.3d at 1100. This black-and-white rule has the benefit of limiting our need to engage in artistic analysis in this context.4
We agree with the district court that the use of Brown’s likeness is artistically relevant to the Madden NFL games. As Brown points out in trying to undermine the status of the games as expressive works, EA prides itself on the extreme realism of the games. As Brown emphasizes in arguing that it is in fact his likeness in the games: “[I]t is axiomatic the '65 Cleveland Browns simply, by definition, cannot be the '65 Cleveland Browns without the players who played for the '65 Cleveland Browns. This fundamental truth applies especially to that team’s most famous player, Jim Brown.” Given the acknowledged centrality of realism to EA’s expressive goal, and the importance of including Brown’s likeness to realistically recreate one of the teams in the game, it is obvious that Brown’s likeness has at least some artistic relevance to EA’s work. The fact that any given version of Madden NFL includes likenesses of thousands of different current and former NFL players does not impact this analysis. In E.S.S., the virtual strip club in question was just one of many virtual structures included by the designers of Grand Theft Auto: San Andreas in an attempt to simulate the feel of East Los Angeles, but we nonetheless concluded that the strip club was artistically relevant to the work. 547 F.3d at 1100. There is no significant distinction to be made here.
Brown questions the artistic relevance of his likeness to' Madden NFL in part by pointing us to the Sixth Circuit’s decision in Parks v. LaFace Records, 329 F.3d 437 (6th Cir.2003). In Parks, civil rights hero Rosa Parks sued the musical duo Outkast under § 43(a) after Outkast released a song called Rosa Parks. Id. at 441. Partially due to the fact that one of the members of Outkast had said that the song was not “intended ... to be about Rosa Parks or the civil rights movement,” the Sixth Circuit concluded that the district court should have at least considered additional evidence before deciding that the use óf Ms. Parks’ name was artistically relevant to the song. Id. at 452-53. Brown alleges that EA has made similar denials of Jim Brown’s relevance to Madden NFL, and thus argues that Brown’s likeness is not artistically relevant to the Madden NFL games. The court in Parks, however, did not rely solely on the band’s denial that the song was about Ms. Parks or the civil rights movement in concluding that there was a factual dispute about artistic relevance. “The composers did not intend [the song] to be about Rosa Parks, and the lyrics are not about Rosa Parks,” the court stated, emphasizing both Outkast’s denials and the court’s own determination that the song’s lyrics were unrelated to Ms. Parks or the civil rights movement. Id. at 452. Here, even if EA’s denials regarding Brown are equivalent to Outkast’s denial regarding Parks, the content of the Mad*1244den NFL games — the simulation of NFL football — is clearly related to Jim Brown, one of the NFL’s all-time greatest players. Moreover, EA’s denials are not equivalent to Outkast’s denial. EA has denied using the aspects of Brown’s likeness that may be protected by the Lanham Act and certain state laws, but such denials are a far cry from Outkast’s outright denial of relevance. In letters to Brown’s attorneys, EA officials have claimed that “Brown has not appeared in any Madden NFL game since 1998,” and that “Brown’s name and likeness does not appear in Madden NFL 08 or any packaging or marketing materials associated with the product.” EA has not denied that Brown’s likeness is relevant to Madden NFL; rather, it has denied that Brown has appeared in the Madden NFL games released since 1998. If the denials are true — that is, if Brown’s likeness does not in fact appear in the games — Brown has no claim at all under the Lanham Act. We do not understand this to be Brown’s position. Outkast’s denial did not similarly undermine Ms. Parks’ Lanham Act claim because Outkast was not denying the use of Parks’ name. In order to have a valid § 43(a) claim based on artistic irrelevance, Brown needs to show both that his likeness was used and that his likeness was artistically irrelevant to the Madden NFL games. If artistic irrelevance can only be proven by accepting the truth of EA’s denial of the use of Brown’s likeness, Brown cannot possibly satisfy both of these burdens. Moreover, in the context of a motion to dismiss, we accept Brown’s factual allegations as true, and Brown alleges that his likeness was used. We must thus assume that EA’s denials are false, meaning they provide no support for artistic irrelevance.5
One of the Sixth Circuit’s animating concerns in Parks was that a celebrity’s name could be “appropriated solely because of the vastly increased marketing power of a product bearing the name of [the celebrity].” 329 F.3d at 454. This is a legitimate concern, but the facts in Parks — specifically, the court’s determination that the lyrics of Outkast’s song may very well have nothing to do with Rosa Parks or the civil rights movement — made that concern much more realistic in that case than in this one. EA did not produce a game called Jim Brown Presents Pinball with no relation to Jim Brown or football beyond the title; it produced a football game featuring likenesses of thousands of current and former NFL players, including Brown. Comparing this case to Parks does not further Brown’s cause.
Brown also asserts that our interpretation of the Rogers test in E.S.S. to require artistic relevance to “merely ... be above zero,” 547 F.3d at 1100, has rendered the Rogers test — described in the Rogers opinion itself as seeking to strike a “balance” between “the public’s interest in free expression” and “protecting] the public *1245against flagrant deception,” 875 F.2d at 999 — an inflexible and mechanical rule that more or less automatically protects expressive works regardless of the deception involved. But a balance need not be designed to find each of the sides weightier with equal frequency. The language in Rogers is clear. “[T]hat balance will normally not support application of the [Lanham] Act unless the [use of the trademark or other identifying material] has no artistic relevance to the underlying work whatsoever____” 875 F.2d at 999 (emphasis added). The Rogers test is applicable when First Amendment rights are at their height — when expressive works are involved — so it is no surprise that the test puts such emphasis on even the slightest artistic relevance. “Intellectual property rights aren’t free: They’re imposed at the expense of future creators and of the public at large,” White v. Samsung Elecs. Am., Inc., 989 F.2d 1512, 1516 (9th Cir.1993) (Kozinski, J., dissenting from denial of rehearing en banc), and the Rogers test applies when this expense is most significant. Our interpretation of the “artistic relevance” prong of the Rogers test in E.S.S. is correct, and Brown fails to allege facts that satisfy that prong in this case.
B
Even if the use of a trademark or other identifying material is artistically relevant to the expressive work, the creator of the expressive work can be subject to a Lanham Act claim if the creator uses the mark or material to “explicitly mislead[] [consumers] as to the source or the content of the work.” Rogers, 875 F.2d at 999. It is key here that the creator must explicitly mislead consumers. “[T]he slight risk that ... use of a celebrity’s name might implicitly suggest endorsement or sponsorship to some people is outweighed by the danger of restricting artistic expression, and [in cases where there is no explicit misleading] the Lanham Act is not applicable.” Id. at 999-1000. This second prong of the Rogers test “points directly at the purpose of trademark law, namely to avoid confusion in the marketplace by allowing a trademark owner to prevent others from duping consumers into buying a product they mistakenly believe is sponsored by the trademark owner.” E.S. S., 547 F.3d at 1100 (internal quotation marks and citation omitted). We must ask “whether the [use of Brown’s likeness] would confuse [Madden NFL ] players into thinking that [Brown] is somehow behind [the games] or that [he] sponsors [EA’s] product,” id., and whether there was an “explicit indication,” “overt claim,” or “explicit misstatement” that caused such consumer confusion, Rogers, 875 F.2d at 1001. Brown puts forth several arguments attempting to show that this second prong of the Rogers test is satisfied, but each of his arguments is unsuccessful.
First, Brown argues that the use of his likeness in the game coupled with a consumer survey demonstrating that a majority of the public believes that identifying marks cannot be included in products without permission at least raises a triable issue of fact as to the second prong of the Rogers test. It is well established that the use of a mark alone is not enough to satisfy this prong of the Rogers test. In MCA, we noted that if the use of a mark alone were sufficient “it would render Rogers a nullity.” 296 F.3d at 902. We reiterated this point in E.S.S., asserting that “the mere use of a trademark alone cannot suffice to make such use explicitly misleading.” 547 F.3d at 1100. Adding survey evidence changes nothing. The test re*1246quires that the use be explicitly misleading to consumers. To be relevant, evidence must relate to the nature of the behavior of the identifying material’s user, not the impact of the use. Even if Brown could offer a survey demonstrating that consumers of the Madden NFL series believed that Brown endorsed the game, that would not support the claim that the use was explicitly misleading to consumers. The Sixth Circuit’s decision in ETW Corp. v. Jireh Publishing, Inc., 332 F.3d 915 (6th Cir.2003), demonstrates this point. In that case, Tiger Woods’ licensing agent, ETW Corporation, brought a Lanham Act claim against the publisher of artwork commemorating Woods’ 1997 victory at The Masters. Id. at 918. A survey was produced in which participants were shown the artwork and asked if they thought Tiger Woods was affiliated or connected with the work or had approved or sponsored it. Id. at 937 & n. 19. Over sixty percent of the participants answered affirmatively, but the Sixth Circuit asserted: “[Plaintiffs survey evidence, even if its validity is assumed, indicates at most that some members of the public would draw the incorrect inference that Woods had some connection with [the work]. The risk of misunderstanding, not engendered by any explicit indication on the face of the [work], is so outweighed by the interest in artistic expression as to preclude application of the [Lanham] Act.” Id. at 937 (footnote omitted). In Rogers itself, the Second Circuit rejected similar survey data for the same reasons. 875 F.2d at 1001. The use of Brown’s likeness together with the cited survey do not provide a valid argument to allow Brown’s case to go forward based on this prong of the Rogers test.
Second, Brown argues that certain written materials that accompanied versions of the game demonstrate EA’s attempts to explicitly mislead consumers about his endorsement or involvement with the game’s production. Unlike mere use of the mark or a consumer survey, statements made in materials accompanying the game are at least the right kind of evidence to show that EA tried to explicitly mislead consumers about its relationship with Brown. Here, however, the statements highlighted by Brown do not show any attempt to mislead consumers. Brown points to materials that say that one of the game’s features was the inclusion of “[f]ifty of the NFL’s greatest players and every All-Madden team.” Since Brown is one of the fifty greatest NFL players of all time and has been named to the “All Madden, All Millennium” team, Brown argues that the statement “explicitly represents that Brown was in EA’s game.” But Brown needs to prove that EA explicitly misled consumers about Brown’s endorsement of the game, not that EA used Brown’s likeness in the game; nothing in EA’s promotion suggests that the fifty NFL players who are members of the All Madden, All Millennium team endorse EA’s game. EA’s statement is true and not misleading.
Third, Brown argues that the changes made to Brown’s likeness for use in certain versions of the game satisfy the second prong of the Rogers test. EA made changes to certain versions of the game that might make a consumer of the game less confident that the player in question was intended to be Brown. Most notably, EA changed the jersey number on the Brown avatar from 32 (the number Brown wore in the NFL) to 37. If these changes had any impact on whether consumers believed that Brown endorsed the game, however, surely they made consum*1247ers less likely to believe that Brown was involved. Brown offers various theories about EA’s legal motives in “scrambling” his likeness for use in the game. It may be true that EA was trying to protect itself from being sued for using Brown’s likeness, under the Lanham Act or otherwise, but an action that could only make consumers less likely to believe that Brown endorsed Madden NFL cannot possibly satisfy the second prong of the Rogers test.
Fourth, Brown cites various comments made by EA officials as evidence that the second prong of the Rogers test is satisfied. As previously discussed, EA attorneys sent letters to Brown’s attorneys stating that “Brown has not appeared in any Madden NFL game since 1998” and that “Brown’s name and likeness does not appear in Madden NFL 08 or any packaging or marketing materials associated with the product.” Brown claims that EA officials contradicted these statements when they allegedly said at a conference held at USC Law School that EA was able to use the images and likenesses of players because it obtained written authorization from both the NFL players and the NFL. The statements made in letters to Brown’s attorneys are irrelevant to this prong of the Rogers analysis. They were not made to consumers, and they do not say anything about Brown’s endorsement of the game. The statement allegedly made at the conference is perhaps the closest Brown comes to offering evidence that EA acted in an explicitly misleading manner as to Brown’s endorsement of the game, but again, the statement was made to a limited audience, not to consumers. If a similar statement appeared on the back cover of a version of Madden NFL, that might satisfy the “explicitly misleading” prong, or at least raise a triable issue of fact, but a statement made at an academic conference about all of the likenesses used in the game could not realistically be expected to confuse consumers as to Brown’s involvement.6
IV
Brown also argues that the district court improperly engaged in factfinding in granting EA’s motion to dismiss. The district court, in Brown’s view, could not possibly have granted the motion to dismiss if it accepted all of the allegations in Brown’s complaint as true, as Brown alleges in his complaint that his likeness is not artistically relevant to Madden NFL and that EA attempted to mislead consumers about his involvement with Madden NFL.
Brown is of course correct that “[o]n a motion to dismiss, the court presumes that the facts alleged by the plaintiff are true.” Halet v. Wend Inv. Co., 672 F.2d 1305, 1309 (9th Cir.1982). We will also “draw[ ] all reasonáble inferences *1248from the complaint in [Brown’s] favor.” Mohamed v. Jeppesen Dataplan, Inc., 614 F.3d 1070, 1073 (9th Cir.2010) (en banc) (internal quotation marks omitted). We are not, however, required to “accept any unreasonable inferences or assume the truth of legal conclusions cast in the form of factual allegations.” Ileto v. Glock Inc., 349 F.3d 1191, 1200 (9th Cir.2003). Brown asserts that there is no artistic relevance and that EA attempted to mislead consumers about Brown’s involvement with Madden NFL, but none of the facts asserted in support of these legal conclusions actually justify the conclusions.
With regard to artistic relevance, even presuming that EA officials have denied the inclusion of Brown’s likeness in the game, the district court could conclude, having reviewed the versions of Madden NFL provided to the court,7 that the likeness of a great NFL player is artistically relevant to a video game that aims to recreate NFL games.
With regard to Brown’s allegation that EA explicitly misled consumers as to his involvement with the game, the factual support Brown offers is simply of the wrong type. Brown would need to demonstrate that EA explicitly misled consumers as to his involvement. Instead, his allegations, if taken as true, only demonstrate that (1) the public can generally be misled about sponsorship when marks are included in products; (2) EA explicitly stated that Brown’s likeness appears in Madden NFL; (3) EA tried to disguise its use of Brown’s likeness, if anything making consumers less likely to believe that he was involved; (4) EA was dishonest with Brown’s attorney about the inclusion of his likeness in the game; and (5) EA suggested to a group of individuals at an academic conference that the players whose likenesses were used in Madden NFL had signed licensing agreements with EA. There is simply no allegation that EA explicitly misled consumers as to Brown’s involvement, and thus no problem with the district court deciding this issue in response to a motion to dismiss.
V
As expressive works, the Madden NFL video games are entitled to the same First Amendment protection as great literature, plays, or books. Brown’s Lanham Act claim is thus subject to the Rogers test, and we agree with the district court that Brown has failed to allege sufficient facts to make out a plausible claim that survives that test. Brown’s likeness is artistically relevant to the games and there are no alleged facts to support the claim that EA explicitly misled consumers as to Brown’s involvement with the games. The Rogers test tells us that, in this case, the public interest in free expression outweighs the public interest in avoiding consumer confusion. The district court’s judgment is thus AFFIRMED.