Written by Eleanor Lackman
In the coming months, the Ninth Circuit en banc will decide Sedlik v. von Drachenberg, a relatively straightforward infringement case that asks whether a tattoo copied from a photograph of Miles Davis was substantially similar in protected expression to the photograph. The jury found that the tattoo was not substantially similar. The admitted copying and the jury’s widely questioned finding of no substantial similarity likely triggered the appellate panel’s two concurrences that all but guaranteed en banc review—in particular, of the second part of the two-part test for substantial similarity that has been applied in the Ninth Circuit for decades.
Multiple amicus briefs and just over an hour of questioning from the en banc judges on September 29, 2026, revealed a wide array of views on the issue but no clear consensus—other than that the Ninth Circuit test is likely to change in some way. Certain themes emerged at the oral argument, and they point to where the Circuit may go.
How did we get here?
Jeff Sedlik photographed Miles Davis in 1989. In 2017, the tattoo artist Kat Von D (real name: Katherine von Drachenberg) traced that photograph, made a stencil, tattooed the image onto a friend’s arm as a gift, and posted the process online; her shop called it “100% exactly the same as the reference.” Sedlik sued. A jury in Los Angeles found the tattoo was not “substantially similar” to the photograph. Sedlik appealed, and in January 2026 a three-judge panel affirmed[EL1] . However, two of the three judges in concurring opinions said the test the jury applied is broken. In June 2026, the full court agreed to rehear the case en banc, and on September 29 eleven judges heard argument about what the test should be.
In deciding the case, the jury was provided the high-level standard embodied in the comment to Instruction 17.19 of the Ninth Circuit Manual of Model Jury Instructions (“Substantial Similarity—Extrinsic Test; Intrinsic Test”)—an instruction that itself does not exist but consists primarily of a comment identifying the standard and directing the court and counsel to craft a custom instruction.
Specifically, the description of the two tests appears in the comment that opens the entry for 17.19:
[A plaintiff] must prove both substantial similarity under the “extrinsic test” and substantial similarity under the “intrinsic test.” The “extrinsic test” is an objective comparison of specific expressive elements. The “intrinsic test” is a subjective comparison that focuses on whether the ordinary, reasonable audience would find the works substantially similar in the total concept and feel of the works.
The extrinsic test is a question of law for the court. If the court determines that the plaintiff has failed to satisfy the extrinsic test, then the infringement claim fails and there is no need to proceed to the intrinsic test. . . . If the extrinsic test is satisfied, the intrinsic test is a matter for the jury.
(Citations omitted.)
Neither the parties nor the court in this case crafted a custom instruction. Instead, the court instructed the jury essentially along the lines of the first paragraph above. It also instructed that the plaintiff had the burden to identify the particular combination, selection, or arrangement of elements in his photograph that he claimed was protected. Because the jury concluded simply that the works were not “substantially similar,” it is not clear whether the plaintiff failed on the extrinsic or intrinsic test, or how the jury applied the standard.
While parties and amici drew stark distinctions between the myriad tests applicable in various circuits nationwide, the main difference between the Ninth Circuit and others is that the Ninth Circuit alone treats the second part of the inquiry as a separate, independently dispositive test, gives that test to the jury, and then treats the jury’s answer as effectively final—the court has never reversed a no-infringement verdict under the intrinsic test, and even the Sedlik three-judge panel declined to do so even while calling the test broken.[1]
The issue is not just vocabulary.
A review of the law shows that every circuit engages in some type of filtering out of non-protectable elements, and ends the inquiry with a holistic comparison—some circuits filtering the unprotectable elements back out at that stage, others not. One problem may be due to terminology: in the Second Circuit, the concept of “total concept and feel” means “do not dissect so finely that you lose the protectable combination overall,” while in the Ninth, “total concept and feel” refers to the entire intrinsic test—applied, at least in this case, with no instruction tying the holistic comparison to the protectable elements, and as a rule with no experts, no dissection, and once a jury has decided it, no review in practice. The Supreme Court has referenced “total concept and overall feel” only in passing, in Petrella v. Metro-Goldwyn-Mayer, Inc., 572 U.S. 663 (2014), quoting an affirmance of a dismissal of a copyright infringement claim at the pleading stage (Peter F. Gaito Architecture, LLC v. Simone Development Corp., 602 F.3d 57, 66 (2d Cir. 2010)).
The divergence more likely derives from local civil procedure.
The divergence likely comes less from doctrine and more from the difference in practice before the Second and Ninth Circuits. Indeed, the phrase on which both circuits now rely originated in a Ninth Circuit case, Roth Greeting Cards v. United Card Co., 429 F.2d 1106 (9th Cir. 1970), where it described the protectable selection and arrangement of otherwise unprotectable elements. However, the Second Circuit has no pattern civil jury instructions adopted by the court (practitioners draw on treatises such as Modern Federal Jury Instructions); the charges are drafted case-by-case and argued every time. The Ninth Circuit, in contrast, has a model civil jury instruction manual that district judges follow closely. So, while the commentary to the instructions (17.19 in particular) does include references to case law, the citations are organized by genre precisely so that courts can craft a work-specific instruction. In practice, instructions tend to track the model’s general language rather than the specific works before the jury, as happened here. Judicial opinions get distinguished, while instructions are read. And some divergence is bound to be the result.
As a result of this structure, the Ninth Circuit’s intrinsic test has led to some heartburn. In the Second Circuit, the court can decide the entire substantial similarity question. In the Ninth, a judge may decide the second half before trial only where the works are so nearly identical that no reasonable juror could find otherwise, and once a jury has decided it, no court has disturbed the result. One judge asked at argument, “How can the court review someone’s feelings?” In New York, the judge can apply the whole test, at any stage, and a verdict can be reviewed. In Los Angeles, that is not so. That precedent would have allowed the district court to decide this case before trial had it found the works near-identical as a matter of law, but the district court instead found triable issues on both tests. And, as some of the judges sitting en banc noted, that ruling—a denied summary judgment motion—became unreviewable once the case was tried to the jury.
Is a change to the test warranted, and if so, what changes might occur?
A large number of amicus briefs advocated for almost as many formulations. Some argued for modification of the intrinsic test, others argued to keep it, others argued that it should be abolished entirely. Some argued that the intrinsic test need not always go to the jury but that as in other circuits, the court could undertake a holistic review as part of its identification and comparison of protectable elements, similar to how the court reviews in the Second Circuit.
At argument, von Drachenberg’s lawyer argued—as any appellee would—that nothing about the test needed to change; von Drachenberg could also rest easily with the knowledge that, as Judge Johnstone observed in his concurrence, the Ninth Circuit has never reversed a no-infringement verdict under the intrinsic test. Sedlik’s counsel criticized the intrinsic test, as well as the tests of other circuits, and, pressed by the court on what should replace the intrinsic test, offered a framework that took shape over the course of the argument.[2] In substance, that framework kept a two-step structure: the court would first decide which elements are protectable (per this framing, a question of fact or a mixed question, in contrast to the jury instruction comment), and if the case survived summary judgment, the jury would be told that the works contain protectable and unprotectable elements, and would compare the works as a whole with that in mind. Counsel indicated that the proposed instruction was linked to 17.19 but, as he acknowledged, a departure from the model—with different guidance depending on the type of works at issue.
Then there is the question of fairness. A plaintiff may say that the current Ninth Circuit test is unfair because a plaintiff has to win twice, while the defendant has to win once—which is true of any claim with more than one element. However, if both parts of the test go to the jury combined into a single question on the verdict form, as happened here, then nobody knows what happened: the conclusion is effectively unreviewable in either direction. A plaintiff cannot undo a no-infringement verdict on the intrinsic test, and a defendant has only the extrinsic prong with which to attack an infringement verdict reached on “feel”—a problem for both sides, especially with an unguided jury that that has been given no framework that a court familiar with the law can later evaluate for legal accuracy.
Of course, the best test is the one that is predictable, consistent, and workable—one that does not leave jurors to decide substantial similarity without a standard they can apply and a court can review. A variable that consists of “a jury will react” and “this could go either way” makes it difficult to evaluate claim risk, which in turn affects errors-and-omissions insurance, indemnity caps, and holdbacks. Even the plaintiff’s lawyer may have trouble taking on a contingency case that she cannot price.
What will change if the test changes?
It was clear from the argument that something likely will change about the Ninth Circuit test. Compounding the court’s task is a pending petition for certiorari to the U.S. Supreme Court in Yonay v. Paramount Pictures Corp., No. 26-61, docketed July 14, 2026, in which the plaintiff-petitioner is challenging the Ninth Circuit’s extrinsic test, the step on which the studio defendant won summary judgment. The response is due on October 16, 2026; Sedlik’s counsel told the en banc court not to wait for it.
The Ninth Circuit’s decision may lean most heavily not into the test itself, but into the standard of review. Practice may change as well: special interrogatories that separate filtration (extrinsic) from the holistic comparison (intrinsic)—and both from the defense of fair use—can help make the review clearer. The en banc court may also remind district courts that the comment to Model Instruction 17.19 already says that the court must tell the jury which elements are unprotectable, what the standard is, and that the jury may consider the work as a whole, to avoid “missing the forest for the trees,” as appellee’s counsel put it, through over-dissection. The comment similarly encourages a more active approach in crafting jury instructions than might occur as to the average model instruction. Judge Johnstone framed it well: “We should defer to juries’ assessments of substantial similarity. But for juries to make those calls consistently, they need more than an empty standard.” The court could adjust the standard, or it could address this issue in part by concluding that the intrinsic finding, like any other jury finding, is subject to ordinary sufficiency review.
At this point, while some change is likely, it is impossible to determine what a change will mean for litigants in the Ninth Circuit. If the intrinsic test is abolished, plaintiffs lose a second hurdle at trial and may win summary judgment more readily in clear cases, while defendants gain something they have never had on this prong: meaningful post-verdict and appellate review. If the test is refined, there may be better guidance, and possibly appellate review. Forum shopping may decrease, or may shift depending on the type of case at issue.
Ultimately, a change in the test may not change the outcome for these parties. While not the primary focus of the rehearing, a remand—whether to the three-judge panel or to the district court—could lead to a retrial, renewed summary judgment motions, or a decision that turns on fair use rather than substantial similarity. For everyone else, however, it is clear that this case has become much bigger than the tattoo. However the en banc court adjudicates the dispute, the opinion will matter more to the next dozens of merits motions on substantial similarity in the Ninth Circuit—and to every party that needs to assess the value and strategy for litigation of such disputes.
[1] The Fourth and Eighth Circuits use the extrinsic/intrinsic vocabulary, but neither has adopted the Ninth Circuit’s rule reserving the intrinsic test to the jury: the Fourth Circuit has declined to adopt it and decides for itself whether a reasonable jury could find intrinsic similarity, while the Eighth Circuit reviews a district judge’s intrinsic finding for clear error.
[2]Sedlik’s petition for rehearing cited the Second Circuit approvingly, but at argument his counsel said that the Second Circuit’s test is “all over the map.”
[EL1]LINK to decision: https://cdn.ca9.uscourts.gov/datastore/opinions/2026/01/02/24-3367.pdf
