Case Information
*1 Before: BATCHELDER, NALBANDIAN, and BLOOMEKATZ, Circuit Judges.
_________________ COUNSEL ARGUED: Mary Massaron, PLUNKETT COONEY, Bloomfield Hills, Michigan, for Appellant. Mark Granzotto, MARK GRANZOTTO, P.C., Berkley, Michigan, for Appellee. ON BRIEF: Mary Massaron, PLUNKETT COONEY, Bloomfield Hills, Michigan, for Appellant. Mark Granzotto, MARK GRANZOTTO, P.C., Berkley, Michigan, for Appellee.
BLOOMEKATZ, J., delivered the opinion of the court in which NALBANDIAN, J., concurred. NALBANDIAN, J. (pp. 18–23), delivered a separate concurring opinion. BATCHELDER, J. (pp. 24–34), delivered a separate dissenting opinion.
*2 _________________
OPINION
_________________
BLOOMEKATZ, Circuit Judge. Aaron Salter spent 15 years in prison for a deadly shooting he did not commit. Salter’s conviction rested on a single eyewitness’s testimony identifying him as one of two shooters. There was no physical evidence or other witness tying Salter to the murder. Instead, the eyewitness identified Salter when the lead investigator, Detective Donald Olsen, showed him a single mug shot of only Salter and said that police had already arrested one of the shooters. That same eyewitness testified that he also identified a different man as the shooter from a photo array. And that man more closely resembled the witness’s description of the shooter than Salter. The investigator’s file also contained a separate closeup photo of that man, suggesting he was a suspect. Salter claims that Detective Olsen failed to disclose any of this information, leading to an unfair trial and his wrongful conviction.
In this civil rights lawsuit, Salter contends that Detective Olsen violated his constitutional rights in two ways: first , by withholding favorable evidence pointing to a different suspect, in violation of Brady v. Maryland , 373 U.S. 83 (1963); and second , by conducting an unnecessarily suggestive identification process that led to the eyewitness’s false identification of Salter. Detective Olsen moved for summary judgment on both these claims, arguing that he was entitled to qualified immunity. The district court denied the motion, and Detective Olsen appealed.
Because we do not have interlocutory jurisdiction over several aspects of this appeal, we dismiss in part. For the parts we review on the merits, we affirm. FACTUAL AND PROCEDURAL HISTORY [1] In the early morning hours of August 6, 2003, Jamar Luster and two of his friends were “hanging out, drinking, and ‘getting high’” on a porch in Detroit, Michigan. Op. & Order, R. 44, PageID 1060–61 (citation omitted). Then gunshots rang out. Luster had an obstructed view from the porch, but he managed to get a view of two shooters 35 to 40 feet away from him. One of the *3 shooters was shorter and fired a pistol, and a taller man fired a long gun. Luster escaped by jumping over the porch banister and crawling to safety, though he was wounded. Tragically, a man standing near the porch died of a gunshot wound.
Rookie homicide detective Donald Olsen responded to the shooting and spoke with Luster in the hospital about four hours later. Luster told Detective Olsen that one of the shooters was “Rob,” a 5'7" Black man in his mid-20s armed with a pistol whom Luster had seen in the neighborhood. Id . at PageID 1062. Luster said he’d never seen the second shooter who fired the long gun but described him as “thin.” Id. Luster also informed Detective Olsen that a man named Earland Collins had shot up the same house about a month before. Luster described the two shooters in a similar way to another police officer a few hours later: (1) “Rob,” a 5'7" Black man in his 20s, weighing between 150 and 170 pounds; and (2) an unknown Black man also in his 20s, standing about 6'0" with a thin build.
Thus far, Aaron Salter had no role in the story. And at 6'4" and 250 pounds, he was much bigger than both suspects Luster described. Yet, with admittedly no articulable basis, Detective Olsen developed a “hunch” that Salter was “Rob.” Id. at PageID 1062–63. Detective Olsen went to Luster’s house that same morning and showed him a single black-and-white mug shot of Salter—that is, not in an array of other suspects. At the time, the Detroit Police Department had a policy that said “[w]itnesses should never be shown only a photograph of the suspect.” DPD Policy, R. 36-20, PageID 935. Detective Olsen also told Luster “the police had picked up the guy with the rifle.” Op. & Order, R. 44, PageID 1063 . Then Luster signed a statement written by Detective Olsen identifying the man in the black-and-white mug shot (Salter) as “Rob.”
At the same meeting, after Luster identified Salter, Detective Olsen showed him a six- person photo array that did not include Salter. Luster later testified that he identified two more individuals as shooters from the photo array, one of whom was Earland Collins, the man Luster said had shot up his house about a month before. Law enforcement records listed Collins at 6'2" and 200 pounds—a close match with Luster’s description of the taller shooter.
*4 Detective Olsen did not disclose Luster’s identification of Collins as a shooter to Salter or his defense attorney. There was also a large closeup photo of Collins, with some attached hand- written notes, that never made it from the police file to the prosecutor or Salter’s lawyer.
Detective Olsen decided to submit the case against Salter to the prosecutor, even though he later explained that he had always thought it “sucked.” Interview Mem., R. 36-18, PageID 930. After all, Luster was the only eyewitness, and no physical evidence or other testimony connected Salter to the shooting. Yet Luster’s identification of Salter based on the black-and- white mug shot apparently persuaded the prosecutor. The prosecutor sought and obtained an arrest warrant for Salter and initiated a murder prosecution against him in Michigan state court.
At a pretrial hearing, Luster reversed course on a few key details. He identified Salter as the taller shooter with “the big gun”—not the shorter shooter named “Rob.” Prelim. Exam Tr., R. 36-9, PageID 814. Luster also said that he’d seen Salter “once or twice” before the shooting, which departed from his earlier statement that he’d never seen the taller shooter before. Id. at PageID 818. For his part, Detective Olsen testified that Luster didn’t identify anyone from the photo array as responsible for the shooting. He also admitted that he could have put Salter’s mug shot in a photo array but inexplicably chose not to. Salter objected to Detective Olsen’s single- person show-up as suggestive and moved to suppress Luster’s identification at the pretrial hearing, but the judge denied his motion.
At trial, Luster and Detective Olsen testified about Luster’s identification of Salter. Luster again identified Salter as the bigger shooter with a “long gun,” even though he first identified Salter as the smaller assailant with a pistol. Trial Tr., R. 36-11, PageID 838–39. He said he never meant to label Salter as the shorter shooter named “Rob,” and that Detective Olsen had misquoted him in recording his identification. Detective Olsen testified that while he showed Luster a photo array without Salter, the array had “nothing to do” with the shooting and was “directed towards other cases.” Trial Tr., R. 36-12, PageID 895. Salter was convicted and sentenced to life without parole. Salter never knew that Luster had identified Collins as a shooter before he was convicted and sentenced.
*5 Salter proclaimed his innocence in a letter to the trial judge in the aftermath of the trial. He also maintained his innocence throughout the post-conviction process, but his appeals and collateral attacks failed. Eventually, one of his attorneys discovered that the closeup photo of Collins had been withheld. An investigation by the Wayne County Prosecutor’s Conviction Integrity Unit followed. The CIU investigation yielded evidence that Collins was probably the taller shooter. When a CIU investigator asked Collins who shot Luster, he responded, “If I tell you, I could get arrested basically.” CIU Investigator Dep., R. 36-24, PageID 995. In his CIU interview, Detective Olsen admitted that the case against Salter “sucked” because it depended on Luster’s shaky identification. Interview Mem., R. 36-18, PageID 930. Detective Olsen commented, “[I] would not be shocked or upset if the case is overturned, and I hope you do something with it.” Id. at PageID 931 (cleaned up). In addition, Luster later recanted his identification of Salter.
Based on the findings of the CIU investigation, the Wayne County prosecutor joined Salter’s attorneys in moving the trial court to vacate Salter’s convictions. Following a stipulation by the parties, the Michigan court that oversaw Salter’s trial vacated the judgment against him and ordered him released from prison. Salter had been incarcerated for 15 years.
Salter then filed this suit under 42 U.S.C. § 1983, alleging that Detective Olsen violated his due process rights by failing to disclose exculpatory evidence and conducting an improper identification process. Detective Olsen moved for summary judgment on both claims based on qualified immunity. The district court denied Detective Olsen’s motions for summary judgment and reconsideration. He timely appealed.
ANALYSIS
Detective Olsen brought this interlocutory appeal, arguing that we should reverse the district court’s decision because: (1) Salter’s claims are barred by the Heck doctrine, (2) Salter has failed to overcome Detective Olsen’s qualified immunity defense, and (3) Salter is collaterally estopped from pursuing a suggestive identification claim. We address these arguments only to the extent our limited jurisdiction permits.
*6 Generally, an order denying summary judgment is not a “final decision” that we have jurisdiction to review under 28 U.S.C. § 1291. Plumhoff v. Rickard , 572 U.S. 765, 771 (2014). There are two exceptions to this general rule that allow for interlocutory appeals from a denial of summary judgment: (1) the collateral-order doctrine, and (2) pendant appellate jurisdiction.
We can review an issue under the collateral-order doctrine if it would be “effectively unreviewable” at the end of the case. Chaney-Snell v. Young , 98 F.4th 699, 708 (6th Cir. 2024) (quoting Mohawk Indus. v. Carpenter , 558 U.S. 100, 106 (2009)). One example is qualified immunity: under the collateral-order doctrine, a defendant can immediately appeal the denial of qualified immunity because summary judgment is the last opportunity to assert immunity from trial. Heeter v. Bowers , 99 F.4th 900, 908 (6th Cir. 2024). But even then, our interlocutory review of qualified immunity denials is limited. We can generally review “purely legal” questions but cannot resolve quarrels with a plaintiff’s record-supported facts. Id. at 909 (citation omitted).
If an issue is not subject to the collateral-order doctrine, it may still be reviewable pursuant to our pendant appellate jurisdiction. For that to be the case, the issue must be “inextricably intertwined” with an appealable one, or its resolution must be necessary to give “meaningful review” to an appealable issue. Chaney-Snell , 98 F.4th at 710 (quoting Swint v. Chambers Cnty. Comm’n , 514 U.S. 35, 51 (1995)).
Our limited jurisdiction does not allow us to reach the merits of much of Detective Olsen’s interlocutory appeal. We now turn to each of his arguments, sifting through those we must dismiss and resolving the hodgepodge of claims that we can review. I. The Heck Doctrine
Detective Olsen argues that Salter’s convictions weren’t properly vacated, so under Heck v. Humphrey , Salter cannot prevail on either of his due process claims. 512 U.S. 477 (1994). The Heck doctrine restricts a plaintiff’s ability to bring a civil rights claim under § 1983 if success on that claim would imply the invalidity of the plaintiff’s conviction or sentence. See Chaney-Snell , 98 F.4th at 707. To pursue a § 1983 claim under those circumstances, the plaintiff must first show that their conviction or sentence has been favorably terminated. See Heck , 512 *7 U.S. at 486–87. That means that the conviction or sentence must have been “reversed on direct appeal, expunged by executive order, declared invalid by a state tribunal authorized to make such determination, or called into question by a federal court’s issuance of a writ of habeas corpus.” Id. at 486–87. Detective Olsen says that Salter hasn’t made that showing.
At this stage, we lack jurisdiction to address Detective Olsen’s Heck argument. In Chaney-Snell v. Young , we held that we have no jurisdiction to review Heck challenges on interlocutory appeal from a denial of qualified immunity. 98 F.4th at 708. Unlike qualified immunity, “the denial of a Heck claim is not ‘effectively unreviewable’ at a suit’s end,” so the collateral-order doctrine does not apply. Id. (quoting Mitchell v. Forsyth , 472 U.S. 511, 527 (1985)). We also lack pendant appellate jurisdiction over Heck challenges because the resolution of qualified immunity “says nothing about the merits” of a Heck claim. Id. at 709–10. Therefore, we do not have jurisdiction over Detective Olsen’s Heck argument, dismiss that part of his appeal, and express no view on its merits. II. Qualified Immunity
Detective Olsen also maintains that the district court improperly denied him qualified immunity on Salter’s constitutional claims. Salter claims that Detective Olsen violated his clearly established Fourteenth Amendment due process rights in two ways: (1) by failing to turn over evidence pointing to another suspect, in violation of Brady ; and (2) by using a single-photo show-up during the criminal investigation. The district court concluded that material disputes of fact precluded qualified immunity on both claims.
We review de novo a district court’s denial of summary judgment on a qualified immunity defense. See Helphenstine v. Lewis County , 60 F.4th 305, 314 (6th Cir. 2023). A public official is entitled to qualified immunity at summary judgment when, viewing the facts in the light most favorable to the plaintiff, the challenged conduct did not violate “clearly established” constitutional rights “of which a reasonable person would have known.” Jackson v. City of Cleveland , 64 F.4th 736, 745 (6th Cir. 2023) (quoting Pearson v. Callahan , 555 U.S. 223, 231 (2009)). The official is entitled to summary judgment unless a “genuine dispute as to any material fact” precludes the defense. Fed. R. Civ. P. 56(a); see Wilkerson v. City of Akron , *8 906 F.3d 477, 481 (6th Cir. 2018). Once the official has asserted qualified immunity, “the plaintiff must show that (1) the official violated his constitutional rights, and (2) at the time of the violation, it was ‘clearly established’ that the officer’s conduct would violate the Constitution.” Heeter , 99 F.4th at 908 (quoting Palma v. Johns , 27 F.4th 419, 428 (6th Cir. 2022)).
We evaluate both the Brady claim and the suggestive identification claim under this standard, to the extent our limited jurisdiction permits. A. Salter’s Brady Claim 1. Brady Violation To defeat qualified immunity on his Brady claim, Salter must first demonstrate that Detective Olsen violated his constitutional rights. The Fourteenth Amendment requires prosecutors, police officers, and forensic scientists to disclose favorable evidence to a criminal defendant that is material to their defense. Brady , 373 U.S. at 87; Clark v. Louisville-Jefferson Cnty. Metro Gov. , --- F.4th ---, 2025 WL 732838, at *8–9 (6th Cir. Mar. 7, 2025). Police officers must turn over evidence with “apparent” exculpatory value to the prosecutor, who must then disclose that evidence to the defendant. Moldowan v. City of Warren , 578 F.3d 351, 381, 388 (6th Cir. 2009) (quoting California v. Trombetta , 467 U.S. 479, 489 (1984)). To sustain a Brady claim based on withheld evidence, Salter must show: (1) the government withheld evidence from Salter, “either willfully or inadvertently”; (2) the evidence favored Salter, “either because it is exculpatory, or because it is impeaching”; and (3) the evidence was material, such that its nondisclosure prejudiced Salter. Jackson v. City of Cleveland , 925 F.3d 793, 814 (6th Cir. 2019) (quoting Strickler v. Greene , 527 U.S. 263, 281–82 (1999)). The district court concluded that Salter satisfied his burden at summary judgment on each of these elements. Detective Olsen disputes all three elements on appeal but raises arguments that we largely have no jurisdiction to review.
On the first Brady element, Detective Olsen argues that he did not withhold evidence pointing to Collins as the shooter. The dissent agrees with him. See Dissenting Op. at 30 n.3. But this is precisely the type of factual dispute that we cannot review at this stage. See Heeter , *9 99 F.4th at 908. Indeed, Detective Olsen abandons any factual argument in his reply brief, recognizing that we don’t have jurisdiction to resolve such factual disputes in an interlocutory qualified immunity appeal. Detective Olsen’s sole legal argument on this element (which we can review) is without merit. He contends that Salter must show that Detective Olsen withheld the evidence in “bad faith.” Appellant Br. at 39. But we have rejected the proposition that an officer’s Brady violation requires “bad faith.” Clark , 2025 WL 732838, at *5 (citing Moldowan , 578 F.3d at 388).
Nor do we have jurisdiction over Detective Olsen’s challenge to the district court’s conclusion on the second and third Brady elements—that a reasonable jury could find that the withheld evidence favored Salter and was material to his defense. In Detective Olsen’s view, and the dissent’s, the withheld evidence was either duplicative of other disclosed evidence or available to Salter from other sources. See Dissenting Op. at 30–32. So it was neither favorable nor material to Salter’s defense. But these arguments present “mixed” questions of law and fact. Clark , 2025 WL 732838, at *6. And our court recently clarified that, in the Brady context, we lack jurisdiction to consider such mixed questions in an interlocutory qualified immunity appeal. See id. at *5–7. We therefore cannot reach the merits of Detective Olsen’s arguments on the remaining Brady elements and dismiss them from his appeal.
2. Clearly Established Law on Withholding Material Evidence Even if he violated Salter’s rights under Brady , Detective Olsen contends that he is still entitled to qualified immunity because, at the time, it was not clearly established that Brady applies to police officers. While we have jurisdiction to review this argument, see id. at *7, we will not consider it because Detective Olsen did not raise it in the district court, see St. Marys Foundry, Inc. v. Emps. Ins. of Wausau , 332 F.3d 989, 995–96 (6th Cir. 2003). Indeed, contrary to his argument now, Detective Olsen conceded below that he had an “obligation to disclose” materially exculpatory evidence to the prosecutor based on Moldowan v. City of Warren , 578 F.3d 351 (6th Cir. 2009). MSJ, R. 29, PageID 380. And, despite the dissent’s contention that Brady did not apply to police officers in 2003, Dissenting Op. at 25 & n.1, “we have held that pre-1990 cases clearly established that the police also have a ‘ Brady -derived’ duty to disclose material exculpatory evidence to the prosecution,” Clark , 2025 WL 732838, at *8. Relying on *10 Detective Olsen’s concession, the district court agreed that Salter satisfied the clearly established prong of his Brady claim. Op. & Order, R. 44, PageID 1088 n.5. Sure, Detective Olsen tried to reverse course in a motion for reconsideration, but the district court was under no obligation to consider his new argument at that stage. See McBride v. Skipper , 76 F.4th 509, 517–18 (6th Cir. 2023). Nor are we, and we reject Detective Olsen’s attempt to reverse course on appeal.
In sum, to the extent we have jurisdiction, we reject Detective Olsen’s challenge to the district court’s denial of qualified immunity on Salter’s Brady claim. The claim may proceed to the jury.
B. Salter’s Suggestive Identification Claim 1. Collateral Estoppel Salter next claims that Detective Olsen violated his due process rights by conducting an unduly suggestive identification by presenting Luster with a single-person show-up consisting only of Salter’s black-and-white mug shot. But before arguing for qualified immunity, Detective Olsen contends that we cannot consider the merits of this claim because Salter is collaterally estopped from pursuing it. In Salter’s criminal case, the state court held that the show-up was not impermissibly suggestive, which Detective Olsen argues precludes Salter from challenging the identification procedure in this lawsuit.
To start, we have jurisdiction over this argument because its resolution is necessary to ensure “meaningful review” of the district court’s denial of qualified immunity. Chaney-Snell , 98 F.4th at 709, 712; see also Peterson v. Heymes , 931 F.3d 546, 553 (6th Cir. 2019); Roberson v. Torres , 770 F.3d 398, 403 (6th Cir. 2014). If Salter is precluded from raising a suggestive identification claim, then he cannot establish a constitutional violation on that basis under the first prong of the qualified immunity analysis. We therefore consider the merits of Detective Olsen’s collateral estoppel argument. On that front, we find his argument unpersuasive.
Federal courts must give a state court judgment the same preclusive effect it would receive in the courts of the rendering state. See Bus. Dev. Corp. of S.C. v. Rutter & Russin, LLC , 37 F.4th 1123, 1129 (6th Cir. 2022). We must therefore look to Michigan law to determine the *11 preclusive effect of the suppression ruling in Salter’s criminal case. Id. Michigan law allows criminal decisions to bind litigants in subsequent civil cases. Peterson , 931 F.3d at 554. But, under Michigan law, a vacated criminal conviction and the interlocutory rulings supporting it have no preclusive effect. See id. at 554–55 (suppression ruling underlying a vacated conviction lacks preclusive effect); Tanner v. Walters , 98 F.4th 726, 735 (6th Cir. 2024). That is true even if the conviction was vacated by stipulation. See Sanford v. Russell , 381 F. Supp. 3d 905, 924 (E.D. Mich. 2019), aff’d , 815 F. App’x 856 (6th Cir. 2020). Here, Salter’s conviction was vacated, so he is not precluded from pursuing his suggestive identification claim against Detective Olsen. [2]
2. Due Process Violation Turning to qualified immunity, we now consider whether Detective Olsen violated Salter’s due process rights by using an unduly suggestive identification procedure. See Haliym v. Mitchell , 492 F.3d 680, 704 (6th Cir. 2007). We analyze suggestive identification claims in two steps. First , we assess whether the identification procedure was “unnecessarily suggestive.” Id. If it was, we consider whether, despite the suggestive procedure, the identification was “nevertheless reliable.” Id. We evaluate these prongs based on the totality of the circumstances. Simmons v. United States , 390 U.S. 377, 383 (1968).
The district court ruled that Detective Olsen abandoned any argument on the first prong of Salter’s suggestive identification claim—whether the identification was unnecessarily suggestive. We agree. Detective Olsen has not disputed that the single-person show-up was unduly suggestive. In his motion for summary judgment, he argued only that Luster “had an independent basis for identification of Mr. Salter apart from the photograph he was shown.” MSJ, R. 29, PageID 387. Put differently, Detective Olsen argued that the identification was reliable, not that it wasn’t unnecessarily suggestive in the first place. And he pursues the same argument on appeal. Accordingly, we proceed to the second element of Salter’s claim—whether the identification was reliable even if the procedure was unnecessarily suggestive.
*12 In assessing reliability, we look at the totality of the circumstances based on five factors: (1) the witness’s opportunity to view the suspect, (2) the witness’s degree of attention, (3) the accuracy of the witness’s prior description of the suspect, (4) the level of certainty demonstrated by the witness at the time he identified the suspect, and (5) the time between the crime and the identification. Manson v. Brathwaite , 432 U.S. 98, 114 (1977). Together, these factors weigh in Salter’s favor.
First , Luster didn’t have a good opportunity to view the shooters. Luster saw them from 35 to 40 feet away, “at night, with limited lighting.” Prelim. Exam Tr., R. 36-9, PageID 814; Interview Mem., R. 36-18, PageID 930. He also didn’t get a good look at the shooters’ faces in the chaos of the moment.
Second , Luster’s attention was compromised. During the criminal proceedings, Luster said he wasn’t drinking or smoking marijuana in the leadup to the shooting, but in this civil litigation, he admitted he had been “smoking weed” and drinking Hennessy. Luster Dep., R. 36- 7, PageID 778, 780. Luster also testified that his priority was to escape from the shooting, not to identify the people around him.
Third , Luster’s description of the shooters bore little resemblance to Salter. Our caselaw strongly indicates that an identification cannot be reliable if there were significant inconsistencies between the witness’s description of the suspect and the person identified. Gregory v. City of Louisville , 444 F.3d 725, 756 (6th Cir. 2006) (citing Webb v. Havener , 549 F.2d 1081, 1086 (6th Cir. 1977)). Luster described a shorter shooter who was 5'7" and thin, and a taller shooter who was 6'0" and also thin. He first said Salter was the shorter shooter, then testified that Salter was the taller one. Either way, neither description matched Salter’s 6'4" 250-pound frame. And, according to the testimony of a Detroit Police Department representative, Detective Olsen would have automatically received Salter’s weight and height when he pulled Salter’s mug shot from the police database and could have seen that they didn’t match.
Fourth , Luster told Detective Olsen that he was “not sure” when he identified Salter as one of the shooters from Salter’s black-and-white mug shot. Luster Dep., R. 36-7, PageID 786. *13 To make matters worse, Luster identified three different people as the two shooters: Salter based on his single mug shot and two people in the photo array, including Collins. This evidence suggests that Luster was not certain about his identification of Salter.
Fifth , timing is the only factor that works in Detective Olsen’s favor. Luster’s identification came only a matter of hours after the incident. That length of time between “observation and identification” does not undermine the reliability of an identification. See Howard v. Bouchard , 405 F.3d 459, 473 (6th Cir. 2005). That said, timing is only one factor, and the other four factors all weigh against reliability here.
Detective Olsen didn’t present an argument on the traditional reliability factors to the district court. Instead, he relied on Luster’s purported familiarity with Salter to argue that the single-person show-up was constitutionally permissible. See Haliym , 492 F.3d at 706 (providing that a court may consider reliability factors that do not fit within the traditional five-factor analysis). And he reiterates that argument on appeal.
Luster’s seemingly limited familiarity with Salter does not outweigh the “corrupting effect” of the single-person show-up. Manson , 432 U.S. at 116. We have held that a witness’s familiarity with a suspect increases the reliability of an identification. Haliym , 492 F.3d at 706. But Luster said he knew “Rob”—who later turned out to be Salter’s cousin—not Salter. That means Detective Olsen did not show Luster a photo of the person Luster claimed to know. In any event, “problems with identification testimony may exist even where the witness is familiar with the [suspect].” Id. Here, multiple problems remain even after we factor in Luster’s alleged familiarity with Salter. Even if he knew Salter, Luster still had to “sufficiently observe” the shooters to identify them as persons he knew. Id. As explained, he did not have the opportunity to do so. Luster was also uncertain about his identification of Salter. And there were glaring discrepancies between Salter’s physical appearance and Luster’s description of the shorter shooter—the one Luster identified as Salter in a signed statement during the show-up. We have “never found that an identification arising from a suggestive format was anything but unreliable when the witness’[s] prior description of the suspect was significantly inconsistent with the suspect’s actual appearance.” Gregory , 444 F.3d at 756. We won’t depart from that course here.
*14 Detective Olsen also exacerbated the unreliability of the single-person show-up by telling Luster that police had already “picked up” the taller shooter—the one Luster later claimed he had identified as Salter from the start. Luster Dep., R. 36-7, PageID 799. [3] The risk of a misidentification increases when “the police indicate to the witness that they have other evidence that . . . the person[] pictured committed the crime.” Simmons , 390 U.S. at 383. And Detective Olsen only increased that risk by showing Luster a single photo of Salter. Gregory , 444 F.3d at 755 (“By presenting only a single suspect to a witness, police convey an implicit message that ‘this is the guy.’”). Under those circumstances, Luster could have felt pressure to make an identification regardless of his degree of confidence, undermining his identification’s reliability.
Detective Olsen further argues that the single-person show-up did not violate Salter’s “core right” to a fair trial, and thus cannot support a due process claim. Appellant Br. at 46. “It is true that an unduly suggestive identification does not, in and of itself, violate constitutional rights.” Gregory , 444 F.3d at 747. But the admission at trial of such an identification, unless harmless, can violate a criminal defendant’s core due process rights. See Webb , 549 F.2d at 1087. Luster signed a statement identifying Salter as a shooter based on the show-up, and that statement was admitted into evidence at Salter’s trial. At trial, Detective Olsen and Luster also testified extensively about Luster’s out-of-court identification. Detective Olsen stated that he showed Luster “one single photograph that related to the shooting,” a “photograph of Aaron Salter.” Trial Tr., R. 36-12, PageID 894. And Luster told the jury that he picked Salter’s photo “[b]ecause he fit the description of the person that was shooting.” Trial Tr., R. 36-11, PageID 861. During the same testimony, Luster also twice identified Salter as his shooter in court. Id. at PageID 839, 883. Because the unreliable single-person show-up appeared in evidence at Salter’s criminal trial, it can form the basis of Salter’s suggestive identification claim. See Gregory , 444 F.3d at 735, 745–47.
*15 3. Clearly Established Law on Single-Person Show-Ups The district court further held that it was clearly established in 2003 that the single-person show-up of Salter was constitutionally impermissible. Detective Olsen disputes this. We agree with the district court.
A constitutional right is “clearly established” if it is “sufficiently clear that every reasonable official would have understood that what he is doing violates that right.” Mullenix v. Luna , 577 U.S. 7, 11–12 (2015) (per curiam) (quoting Reichle v. Howards , 566 U.S. 658, 664 (2012)). To show that a right is clearly established, a plaintiff need not provide “a case directly on point, but existing precedent must have placed the . . . constitutional question beyond debate.” Id. (quoting Ashcroft v. al-Kidd , 563 U.S. 731, 741 (2011)). The dispositive question is whether the official had notice that his “ particular conduct” violated the Constitution. Id. (quoting al- Kidd , 563 U.S. at 742).
In 1967, the Supreme Court warned that “showing suspects singly to persons for the purpose of identification has been widely condemned.” Stovall v. Denno , 388 U.S. 293, 302 (1967). It allowed a single-person show-up because it was the only possible identification procedure during an exigency and explained that courts must evaluate “the totality of the circumstances” to determine whether a single-person identification procedure amounts to a violation. Id.
In 1977, our court applied Stovall and its progeny in Webb v. Havener . 549 F.2d at 1083–85. In Webb , two robbery witnesses described the robbers to the police “in general terms.” Id. at 1082–83. The police told the witnesses to wait at the station while they went to get a suspect, and then they presented Webb to the witnesses without arranging for a lineup. Id. at 1086. The witnesses identified Webb as one of the robbers, and Webb was convicted of armed robbery with the witnesses’ identification as the sole evidence connecting him to the crime. Id.
We held that the single-person show-up of Webb was unduly suggestive, and that a subsequent in-court identification was not harmless error because it lacked “an independent basis.” Id. at 1086–87. We explained that police officers cannot focus a witness’s attention on one suspect without an articulable need for employing such a procedure. Id. Yet the officers in *16 Webb offered “[n]o explanation” for why they failed to arrange a lineup, and there was no reason for a “hurried” process as in Stovall . Id. at 1086. We also held that several reliability factors weighed in Webb’s favor: the witnesses had observed the robbers for only “a couple of minutes” while at gunpoint; their description of the robbers didn’t match Webb’s appearance; and their testimony was at times inconsistent. Id. The admission of the overly suggestive identification at trial therefore violated Webb’s due process rights. Id. at 1086–87; see also Gregory , 444 F.3d at 746–47 (denying qualified immunity for a single-person show-up that occurred in 1992).
Detective Olsen’s decision to present Luster with a single black-and-white mug shot of Salter was clearly unlawful for the same reasons. Detective Olsen admitted that nothing prevented him from presenting Salter’s photo as part of an array. Adding to the suggestive nature of the identification, before showing Salter’s photo to Luster, Detective Olsen told Luster that the police had apprehended one of the shooters. That statement could have further influenced Salter’s identification by suggesting that Detective Olsen had an independent basis to suspect Salter (he did not). Luster’s identification also had the same indicia of unreliability as the one in Webb : Luster saw the shooters briefly at night while fleeing from gunfire; his description of the shooters’ weight and height did not match Salter’s; and his identification was inconsistent—he identified three people as the two shooters.
The dissent counters that Salter’s rights could not have been clearly established, otherwise his attorney would have objected more forcefully to evidence of the show-up during trial. Dissenting Op. at 27–28. (The trial court, however, had already denied Salter’s pre-trial motion to suppress the show-up as suggestive.) Alternatively, the dissent says the prosecutor would not have relied so heavily on the show-up in seeking an indictment if it were clearly unconstitutional, and the state court would have excluded it. Id. at 32–33. But our inquiry is not based on the actions of Salter’s lawyer, the prosecutor, or the now-vacated rulings in the state court criminal proceedings. Based on binding precedent, a reasonable officer would have known *17 that, under these circumstances, the single-person show-up would not produce a reliable identification. So its admission at trial violated Salter’s due process rights. [4]
What’s more, at the time, the Detroit Police Department’s policy also prohibited the show-up, stating that “[w]itnesses should never be shown only a photograph of the suspect.” DPD Policy, R. 36-20, PageID 935. That policy is of course not dispositive of our qualified immunity inquiry, but it is further proof that Detective Olsen was on notice that his actions were unlawful. See Martin v. City of Broadview Heights , 712 F.3d 951, 962 (6th Cir. 2013). We therefore hold that Detective Olsen was not entitled to qualified immunity on Salter’s suggestive identification claim.
CONCLUSION
For these reasons, we affirm in part and dismiss in part for lack of jurisdiction. *18 _________________
CONCURRENCE
_________________
NALBANDIAN, Circuit Judge, concurring. I join the majority opinion in full. But I write separately to note two areas where I question this court’s precedents. First, I believe that recent Supreme Court precedent fatally undermines 42 U.S.C. § 1983 suits based on unduly suggestive identification procedures. These protections, like Miranda , are prophylactic rules that do not create a right to sue under § 1983. Second, I harbor reservations about this court’s imposition of individual liability on police officers for derivative Brady violations. Grafting these Brady -derived claims onto our constitutional tort jurisprudence has undermined the effectiveness of our review of qualified immunity.
I.
I start with some background. We first allowed a § 1983 suit against a police officer who conducted an allegedly suggestive identification in Gregory v. City of Louisville , 444 F.3d 725, 745–47 (6th Cir. 2006). There, the police officer asked Gregory “to agree to a one-on-one show- up with” a rape victim and asked him to “repeat the words uttered by [the victim’s] rapist.” Id. at 733. After being first convicted and then exonerated in state court, Gregory sued the officer under § 1983. Id. at 735. The district court denied qualified immunity against Gregory’s claim that the officer “used an unduly suggestive show-up procedure in getting [the victim] to identify [him].” Id. at 745.
On appeal, the panel concluded that the plaintiff had alleged “the violation of a known constitutional right”—namely, the “right to be free of unduly suggestive identification procedures.” Id. at 746–47. The panel rejected the officer’s argument that the prosecution’s use of the identification at trial was the actual violation. So it denied qualified immunity, concluding that the officer could not escape “liability if he reasonably should have known that use of the identification would lead to a violation of Plaintiff’s right to a fair trial.” Id. at 747.
Even when it was decided, Gregory rested on shaky foundations. First, Gregory said that “the Supreme Court has held that police officers must evaluate the totality of the circumstances *19 and reach a reasoned conclusion as to whether an identification procedure is impermissibly suggestive or not.” Id. at 746 (emphasis added) (citing Neil v. Biggers , 409 U.S. 188, 199–200 (1972)) . But the Supreme Court has only required courts to conduct such an inquiry to determine whether an identification is reliable despite being tainted by a suggestive procedure. See Perry v. New Hampshire , 565 U.S. 228, 239 (2012) (citing Biggers , 409 U.S. at 201). Nothing in Biggers imposed a similar requirement upon investigating officers. So the panel decision dramatically expanded Biggers ’s scope.
Second, the panel suggested that this “[c]ourt has entertained allegations of suggestive identification procedures as viable constitutional tort claims.” Gregory , 444 F.3d at 746 (citing Hutsell v. Sayre , 5 F.3d 996, 1005 (6th Cir. 1993)). But there we explicitly rejected a claim that a police officer had violated a plaintiff’s due-process rights by conducting an allegedly suggestive line-up, even though the identification was “presumably used at trial.” Hutsell , 5 F.3d at 1005. And we stated that absent “extraordinary circumstances, such as coercion or other police misconduct,” allegedly suggestive identifications could not give rise to a cause of action under § 1983. Id. So Hutsell hardly supports the viability of these claims. Still, Gregory is the law in this Circuit.
But after the Supreme Court’s recent decision in Vega v. Tekoh , 142 S. Ct. 2095 (2022), its continued viability is questionable. In Vega , the Supreme Court held that plaintiffs may not bring § 1983 suits against police officers “based on the allegedly improper admission of an ‘un- Mirandized ’ statement in a criminal prosecution.” Id. at 2099. The Court clarified that Miranda created “prophylactic rules” to safeguard the right against the introduction of compelled, self- incriminating statements in a criminal proceeding. Id. at 2101. And the opinion stressed that Miranda did not hold that a violation of its rule “necessarily constitute[s] a Fifth Amendment violation.” Id. at 2101. So failure to follow Miranda is not a constitutional violation that gives rise to a claim under § 1983. Id. at 2106, 2108.
Vega ’s reasoning applies to our suggestive identification cases. As with Miranda violations, the Supreme Court has stated that “a suggestive preindictment identification procedure does not in itself intrude upon a constitutionally protected interest.” Manson v. Brathwaite , 432 U.S. 98, 113 n.13 (1977). Thus, like Miranda , the judicially created protection *20 against suggestive identification procedures is only a “prophylactic rule.” Hutsell , 5 F.3d at 1005. Other circuits have agreed. See, e.g. , Alexander v. City of South Bend , 433 F.3d 550, 555 (7th Cir. 2006); Pace v. City of Des Moines , 201 F.3d 1050, 1055 (8th Cir. 2000). While it is designed to safeguard a constitutional interest, violation of this rule doesn’t necessarily violate the Sixth Amendment’s guarantee of a fair trial or the Fourteenth Amendment’s due-process requirement. So an impermissibly suggestive identification procedure is not itself an independent constitutional deprivation that would support a § 1983 claim.
Although I recognize that Gregory ties this panel’s hands, I am persuaded that Vega casts doubt on our precedent. The protection against unduly suggestive identification procedures is, like Miranda , only a prophylactic rule and should not be actionable under § 1983. But Olsen hasn’t made this argument, so it lies outside the scope of our review.
II.
The Brady claim—or more precisely, the Brady -derived claim—doesn’t pose the same problem. A Brady violation is a constitutional deprivation because the prosecutor’s failure to disclose material information violates due process. 373 U.S. 83 (1963). So Brady is not a prophylactic rule. Importantly, under our caselaw, a police officer has an “analogous or derivative” Brady obligation. Moldowan v. City of Warren , 578 F.3d 351, 379–81 (6 th Cir. 2009). This analysis hinges on the officer “suppressing exculpatory evidence.” Id. at 382. So an officer’s failure to disclose information must have caused an unconstitutional suppression. And the derivative obligation is inextricable from the constitutional deprivation. But these claims have different problems.
A.
To begin, I question whether the law was clearly established as of the events in this case. Though this court held that Brady -derived claims were clearly established as early as 1990, this conclusion rested only on persuasive authority from other circuits. Id. at 382. It’s true that the Supreme Court has suggested that “absent controlling authority,” clearly established law can be supported by “a robust ‘consensus of cases of persuasive authority.’” Ashcroft v. al-Kidd , 563 U.S. 731, 741–42 (2011) (quoting Wilson v. Layne , 526 U.S. 603, 617 (1999)). But Moldowan *21 only cited cases from three circuit courts holding that the police’s Brady -derived obligations were clearly established by 1990. 578 F.3d at 382. To my mind, three circuits doesn’t represent a consensus, much less a robust one. So these citations don’t show that by 1990 existing precedent had placed the “constitutional question beyond debate.” al-Kidd , 563 U.S. at 741. Still, Moldowan continues to bind this panel.
B.
Next, I also share the concerns expressed by Judge Murphy in his recent concurrence in Clark v. Louisville-Jefferson County , No. 24-5061, 2025 WL 732838, at *1 (6th Cir. March 7, 2025). Namely, whether a plaintiff should have to show bad faith on the part of the officer and whether we have jurisdiction to review Brady ’s materiality element as part of the qualified immunity inquiry. Id. at *11 (Murphy, J., concurring). The Clark decision tells us that our precedents require us to answer no as to both questions. Id. at *6–9 (majority opinion). But I am unconvinced that this is correct as a matter of law.
Regardless, the current state of our caselaw artificially limits the scope of our review in these cases when there is any question of materiality. This limitation makes little sense in § 1983 cases based on Brady —especially when we are talking about liability for police officers and not prosecutors. And I believe the issue stems from the chimerical nature of the Brady -derived claims recognized by Moldowan .
Under the Brady framework, “an inadvertent nondisclosure has the same impact on the fairness of the proceedings as deliberate concealment.” Stickler v. Greene , 527 U.S. 263, 288 (1999). Thus, the culpability of the prosecutor is not at issue. A prosecutor can violate Brady even if he acted in good faith or lacked any knowledge of the material evidence that was suppressed. And in most instances of a Brady violation, the remedy for defendants is a new trial—not damages. For this reason, claims under Brady have always fallen “within the traditional core of habeas corpus and outside the province of § 1983.” Skinner v. Switzer , 562 U.S. 521, 536 (2011). So Brady ’s familiar three-step framework of favorability, suppression, and materiality developed to ensure the fairness of a trial—not as a mechanism to punish prosecutorial misconduct. And the Brady doctrine’s lack of a knowledge or intent requirement *22 as to the suppression is softened by the fact that prosecutors are absolutely immune “from civil liability for the non-disclosure of material exculpatory evidence at trial.” Koubriti v. Convertino , 593 F.3d 459, 470 (6th Cir. 2010).
But unlike prosecutors, police officers don’t have absolute immunity. They must make do with the more ad hoc doctrine of qualified immunity. See King v. Harwood , 852 F.3d 568, 584 (6th Cir. 2017). And because qualified immunity is “not just a defense against a damages award but also an ‘ immunity from suit ’” an officer can appeal the denial of qualified immunity under the collateral-order doctrine. Chaney-Snell v. Young , 98 F.4th 699, 708 (6th Cir. 2024) (quoting Mitchell v. Forsyth , 472 U.S. 511, 526 (1985)). Without immediate appellate review, the immunity “is effectively lost if a case is erroneously permitted to go to trial.” Himmelreich v. Fed. Bureau of Prisons , 5 F.4th 653, 662 (6th Cir. 2021) (quoting Mitchell , 472 U.S. at 526). And much of our § 1983 jurisprudence has developed to match the contours of the collateral- order doctrine, which limits our review to resolving questions of law.
So putting it all together, Brady ’s requirements of favorability and materiality developed free from the constraints imposed by the collateral-order doctrine. But this court’s current practice of allowing Brady -derived claims against officers, combined with a limitation on what we can review, pits this doctrine against the limitations of the collateral-order rule. And the officer’s derivative Brady obligation only exists “[w]here the exculpatory value of a piece of evidence is apparent.” Moldowan , 578 F.3d at 388 (internal quotation marks omitted). So the constitutional violation depends not just on favorability and materiality, but also on the added question of whether the exculpatory value was apparent. But—for better or worse—we have concluded that “materiality under Brady is a mixed question of fact and law for the jury” and that “this Court lacks jurisdiction to entertain [such appeals.]” Gregory , 444 F.3d at 744. These competing requirements leave us in a jurisdictional bind. We cannot review the legal question of whether there was a constitutional violation because we lack jurisdiction over materiality.
This jurisprudential wrinkle has created real world problems. As Judge Murphy points out, such treatment of materiality is out of step with how we treat other mixed questions of law and fact in the § 1983 context. Clark , 2025 WL 732838, at *12 (Murphy, J., concurring) (citing Williams v. Mehra , 186 F.3d 685, 690 (6th Cir. 1999)). And I believe that this discord springs *23 from the fact that we have superimposed § 1983 liability onto Brady —a trial right focused on the actions of prosecutors that does not account for the practicalities of constitutional tort litigation. Since the materiality of suppressed evidence is the key to Brady , we are effectively barred from resolving the central legal question of whether—based on undisputed facts—a constitutional violation occurred.
So we are left with two warring doctrines. A review of materiality is essential to determine whether the plaintiff has alleged a constitutional violation. But our caselaw prevents us from reaching that question on jurisdictional grounds. Thus, it insulates the district court’s decision from review at the only time that matters for qualified immunity—that is, before the official has endured a suit. If we truly lack jurisdiction to hear such appeals, our review of denials of qualified immunity in Brady -derived claims is rendered toothless. And qualified immunity’s protection from suit becomes a hollow promise. While we cannot resolve this quandary today, I am unsure how much longer we can avoid it.
*24 _________________
DISSENT
_________________ ALICE M. BATCHELDER, Circuit Judge, dissenting. The undisputed facts in this case matter. It is deplorable that Aaron Salter spent 15 years in prison for a murder he did not commit. And there are people to blame, but if Detective Donald Olsen is one of them, he is way down the list. Certainly, Jamar Luster is at the top of that list. As everyone concedes, Luster looked at Salter’s photo on the day after the shooting and said Salter was the shooter. At the preliminary examination, Luster looked at Salter in open court, face to face, and—under oath and subject to cross-examination—insisted emphatically and unequivocally that Salter was the shooter. At trial, Luster’s testimony convicted Salter: Luster was the State’s only eyewitness to the shooting and once again, under oath and subject to cross-examination, Luster testified emphatically, unequivocally, and repeatedly that Salter was the shooter. It would be 15 years before Luster would recant this testimony and tell a new story about what he said and did back in 2003.
But Salter also has himself to blame, followed closely by his trial and appellate counsel. Well before his trial, Salter learned from both his cousin and an inmate he met in prison that E Collins was the shooter, and Salter realized that Luster had misidentified him as E Collins, or had falsely accused him. Either Salter did not reveal this information to his trial counsel or, for reasons unknown, his counsel chose not to introduce this evidence in Salter’s defense at trial. But immediately after trial, Salter wrote a letter to the trial court judge, explaining this in detail. The judge entered that letter into the record, meaning that Salter’s appellate counsel had access to it for Salter’s appeal, but again for reasons that are unknown and inconceivable, Salter’s appellate counsel did not raise that on appeal, even as a claim of ineffective assistance of trial counsel.
Finally, Salter could reasonably blame the prosecutor, who prosecuted the case based on a single witness’s allegedly improper photo identification; or the trial court judge who refused defense counsel’s motion to suppress the photo identification; or the appellate judges who rejected the claim on appeal that the unduly suggestive photo identification warranted a new *25 trial, holding instead that Luster’s in-court identification cured any harm from the photo identification.
But Salter does not blame any of those people. Salter and the lead opinion put all of the blame on Det. Olsen. Neither the record nor the law supports that. There is no Brady evidence in this record and, therefore, no Brady violation even if Brady applied boundlessly to police officers back in 2003, which it did not. [1] Nor did Det. Olsen knowingly violate clearly established law, circa 2003, when he used the single photo to identify Salter. Qualified immunity protects all but the plainly incompetent and those who knowingly violate the law. Nothing in this record suggests that Det. Olsen was plainly incompetent or knowingly violated the law. Because I disagree with the lead opinion on its reading of both the record and the law, I must respectfully dissent.
I.
Here is what we know now, 20 years later. At a little after 1 a.m. on August 6, 2003, E Collins and another man shot and killed Willie Thomas. Jamar Luster was one of three witnesses caught in the crossfire. At the hospital, Luster told Det. Olsen that one of the shooters was a man named “Rob,” whom he had seen waiving a gun earlier that day, but he had never seen the other shooter before. The next day, Olsen showed Luster a mugshot photo of Salter’s face, and Luster identified him as “Rob.” Olsen then showed Luster a six-person photo array, which included a photo of E Collins. Luster later (at a deposition in 2019) claimed that he picked E Collins from that photo array and told Olsen that Collins was one of the shooters. But, according to Olsen, Luster did not identify any other shooter from the photo array. Either way, we know that Olsen did not pursue anyone else as a shooter and, specifically, he did not *26 investigate or arrest E Collins. Olsen turned the case over to the prosecutor, who charged Salter with Thomas’s murder.
But Salter was not “Rob,” nor was he one of the shooters. “Rob” was Salter’s cousin, Robert Clark—the “Rob” whom Luster saw waiving a gun earlier that day. And Salter had an alibi and at least three alibi witnesses to prove that he was elsewhere at the time of the murder.
At the preliminary examination on September 22, 2003 (seven weeks after the shooting), Luster testified and identified Salter in court as one of the two shooters. Confronted with his prior description of “Rob” as being 5'7", 150 lbs., and light complected, whereas Salter was clearly 6'4", 250 lbs., and dark complected, Luster changed course and said Salter was not “Rob”; “Rob” was the other, smaller shooter. Luster testified that he told Olsen that one shooter was named “Rob,” and that “I told him another name too.” But Salter’s attorney did not ask Luster to give him that other name, he just moved on, leaving it unspoken and unrecorded. Luster also testified that he picked two people from the photo array as shooters, stating “him [meaning Salter] and the other guy, Rob, or whatever.” But Salter was not pictured in the photo array. Salter’s attorney did not ask Luster to show him which of the two photos he had picked, so that also remains unknown. When Olsen testified, Salter’s attorney asked whether Luster had identified anyone in the photo array and Olsen said: “Not as the persons that were responsible for the shooting, but just I [sic] wanted to find out if he knew any of those people in that picture.” At the end of that preliminary examination, Salter’s attorney moved to suppress Luster’s pretrial photo identification of Salter, arguing that Olsen should have used only a photo array, and not the mugshot, for Luster’s identification of Salter. The State argued that Luster “said he knew him [Salter]. He’d seen him in the past. He picked him out and he told [Olsen] that he had seen him in the past and seen him in the neighborhood. He described him and said that’s him. [Salter]’s not in custody. His rights were not violated at all.” The trial court denied the defense motion, holding that “there is no constitutional requirement to have [the identification] done in a certain way.”
About a week later, on September 29 or 30, 2003, Salter’s cousin Rob told him that E Collins had committed the murder and “not to worry” about the murder charge, that he would not be convicted because he (Salter) did not do it. Either Salter did not tell his attorney about *27 this information, or his attorney chose not to use it, because this information was not introduced at Salter’s trial. Nor was the evidence of his alibi defense.
At that jury trial, in December 2003, Luster was the State’s key witness and immediately and unequivocally identified Salter in open court as one of the shooters. On direct examination, Luster said that, from the photos shown to him by Det. Olsen, he had picked three people who looked like the shooters. Luster repeated this on cross examination, insisting four separate times that he had picked three people from the photo array that Olsen showed him. But Salter’s attorney never asked Luster to re-identify any of those photos, so which three of the six Luster claimed to have picked remains unknown. [2] Luster might have—certainly could have—identified E Collins, whose photo was number six in that photo array, and who Luster now claims to have identified for Det. Olsen from that photo array. But Luster did not identify or name E Collins at Salter’s trial. The only person Luster did identify at trial was Salter, and Luster identified him several times. Luster admitted that he signed a statement identifying Salter as “Rob,” but then said that was a “mistake”; he did not identify Salter as “Rob,” because “Rob is way smaller than [Salter].”
Det. Olsen testified that Luster identified Salter as one of the shooters from the mugshot, but denied that Luster had identified anyone else in the photo array as having any connection with the shooting. When Salter’s attorney asked Olsen on cross-examination about his showing Luster the mugshot rather than using a photo array, the State argued that a “photo array is not mandated by law” unless the suspect is “in custody,” which Salter was not. On re-direct, the State asked Olsen: “As it relates to the photo array, there was nothing legally mandating you to do a photo array,” to which Olsen agreed there was not. On re-cross, Salter’s attorney pressed Olsen as to whether the mugshot was unduly suggestive:
*28 Salter’s attorney: Don’t you think it would have been a little less suggestive had you put Mr. Salter in an array with five other photos as opposed to just showing Mr. Salter’s [mugshot] photo to Mr. Luster just by itself, do you think that would have been a little less suggestive?
Det. Olsen: No. Not that in this case. Salter’s attorney: Okay. That’s all I have. This is certainly evidence, arguably dispositive evidence, as to the “clearly established
law,” circa 2003, about an allegedly “unduly suggestive” identification in this case. Consider it this way. If the law were actually as clearly established as the lead opinion would have us believe, then why didn’t Salter’s attorney object and say something like: Wait a minute, that mug-shot identification violates Salter’s constitutional rights, and everybody here knows it .
Salter did not testify at trial, nor did his defense attorney present any affirmative defense. The jury convicted Salter of first degree murder. Within days of that verdict, Salter wrote a letter to the trial judge stating his version of the events on August 5 and 6, that “Rob” was his cousin, and that he (Salter) had an alibi. Salter later attested at a deposition to the truth of that letter. As relevant here, Salter wrote: “I happen to see Rob on Sept 29 or 30 [and] he told me not to worry because he know it couldn’t be me because ‘E’ did it.” Also: “One guy [in prison] name William Taylor said ‘E’ personally told him he did that crime.” And: “The reason Jamar Luster pick me is unknown to me. ‘E’ was in the lineup & Jamar previously statement he stated someone named ‘E’ shoot this house up before.” The critical takeaway from this letter is that, at the time of his trial, Salter knew, and had witnesses, that E Collins was the shooter, and that Luster had misidentified him (Salter) as Collins. The trial judge entered this letter into the record, but Salter’s appeal did not refer to any of this information.
On direct appeal, Salter’s appellate counsel raised only one issue, ineffective assistance of trial counsel for the failure to exclude Luster’s in-court identification on the basis that “the pretrial identification procedure was improperly suggestive” “because investigators showed Luster only a single photograph.” The Michigan appellate court rejected that claim, explaining:
[I]t is likely that the trial court would have excluded testimony concerning Luster’s pretrial identification of [Salter] . . . however, it is also likely that the trial *29 court would have nonetheless allowed Luster’s in-court identification of [Salter] as having an independent basis. Indeed, Luster first identified [Salter] only a few hours after the shooting, and testified that he clearly observed [Salter], whom he already knew from the neighborhood, during the shooting. Consequently, we find that [Salter] has failed to establish any prejudice resulting from his trial counsel’s failure to seek the exclusion of Luster’s identifications.
Michigan v. Salter , No. 253401, 2005 WL 954961, at *1-2 (Mich. Ct. App. Apr. 26, 2005). So, the Michigan appellate court, like the trial court, rejected Salter’s claim that Olsen’s use of the single-photo mugshot for the identification necessarily violated his constitutional rights.
In his M.C.R. § 6508 post-conviction motion (in 2010), Salter claimed ineffective assistance of trial counsel for the failure to present his alibi defense and witnesses. The court rejected the claim as trial strategy, and rejected the claim regarding counsel’s failure to raise the issue on appeal. Seven years later, Salter filed a second M.C.R. § 6508 motion, asserting two pieces of new evidence: a polygraph test of Salter, which he passed, and an unsworn, unnotarized statement from E Collins asserting that Salter was not the shooter, but rather “DeVaughn Porter” was the shooter. The court said that was not enough and denied the motion.
In August 2018, about 15 years after Salter’s conviction, attorneys from the Conviction Integrity Unit (CIU) of the Wayne County (Michigan) Prosecutor’s Office and attorneys from the Federal Defenders’ Office determined that the conviction was based on “mistaken identification by the main witness in the case,” and agreed to dismiss the criminal charges and set aside Salter’s conviction. The state trial court accepted the stipulation and vacated Salter’s conviction.
So, Salter sued Det. Olsen (who is now retired) under § 1983, for a violation of his constitutional rights, and demanded $75 million. Olsen claimed qualified immunity, which the district court denied as to two of the claims: (1) an alleged Brady violation based on Olsen’s alleged failure to include a larger picture of E Collins in the file that he gave to the prosecutor, and (2) the use of the single photo of Salter to obtain the identification from Luster, which Salter claimed was unduly suggestive. This is an interlocutory appeal from the denial of qualified immunity.
II.
*30 Det. Olsen argues that evidence is not Brady material if the defendant and his lawyer already knew about it, and the fact was available to them from another source. See Henness v. Bagley , 644 F.3d 308, 325 (6th Cir. 2011) (“Since Henness was aware of the essential facts that would enable him to take advantage of the exculpatory evidence, no Brady violation occurred.”); Coleman v. Mitchell , 268 F.3d 417, 438 (6th Cir. 2001) (“The Brady rule does not assist a defendant who is aware of essential facts that would allow him to take advantage of the exculpatory evidence at issue.”); United States v. Todd , 920 F.2d 399, 405 (6th Cir. 1990).
Olsen points out that Salter knew (from his cousin Rob and others) that E Collins was the shooter; knew that Luster had identified at least two (as he said at the preliminary examination) or even three (as he said at trial) suspects from the six-person photo array; knew that E Collins was pictured in the photo array whereas he (Salter) was not; and knew that—for reasons “unknown to” Salter—Luster had misidentified him. Therefore, Olsen has shown that Salter and his counsel “were aware of the essential facts” that allowed them to take advantage of the exculpatory evidence (i.e., that Collins was the shooter), so the larger picture of Collins was not Brady material. [3]
For his part, Salter sidesteps whether the larger picture, alone, is Brady material. He urges a Brady violation based on two “essential facts” working together—two “new facts” that he says were unknown to his attorney at trial because Olsen did not provide them to the prosecutor, and that are known to him now only because Luster has offered new testimony in 2019 and 2020. Fact One: Luster attested that if someone (e.g., Salter’s trial attorney) had shown him the other, “larger photo of Collins, he would have identified [E Collins] as the taller shooter with the rifle.” R. 36-25, Luster affidavit, Mar. 16, 2020. Fact Two: Luster testified that he *31 identified E Collins to Olsen in 2003, i.e., he “picked ‘E’ [Collins] out of the photo array as a shooter and told Olsen that ‘E’ was one of the shooters.” R. 36-7, PgID 786, Luster depo., Apr. 26, 2019.
The problems with Luster’s two statements are immediately apparent. Conceptually, they contradict each other: in the first, Luster says that if he had been shown the larger photo, he would have identified Collins; in the second, he says that he did identify Collins from the smaller photo-array photo. If the second statement (Fact Two) is true, then it renders the first statement meaningless: i.e., because Luster had already identified Collins as a shooter from the photo array, the larger photo was not needed for Luster to identify him. And Fact Two is not “newly discovered” evidence as we ordinarily use that term (i.e., evidence that could not have been discovered through the exercise of reasonable due diligence), given Luster’s testimony at the preliminary examination that he had identified another shooter to Olsen by name, and his trial testimony that he had identified three others to Olsen from the photo array. Had Salter’s attorney simply asked Luster, “Who?”, he would have obtained Fact Two back then, i.e., that Luster had identified Collins. Also, given that Det. Olsen testified at both the preliminary examination and again at trial, and specifically said that Luster did not identify anyone from the photo array, this Fact Two statement does not constitute Brady material. Believing Luster and disbelieving Olsen makes this a claim that Olsen committed perjury, not that he withheld potentially useful evidence (i.e., the larger photo of E Collins) prior to trial.
Regardless, even if we accept that Luster pointed to E Collins in the photo array and told Olsen that Collins was one of the shooters (Fact Two), and assume that Olsen withheld this identification from Salter’s trial attorney, it would not violate Brady in these circumstances. “[T]here is no Brady violation if the defendant knew or should have known the essential facts permitting him to take advantage of the information in question, or if the information was available to him from another source.” United States v. Graham , 484 F.3d 413, 417 (6th Cir. 2007) (quotation marks and citation omitted). Given Luster’s repeated insistence during his preliminary examination and trial testimony that he had identified others from the photo array, Salter’s attorney certainly had the essential facts necessary to show Luster the photo array and ask him to point out the others he had identified to Det. Olsen, whereupon Luster—according to *32 his depositions later—would have identified E Collins. Salter’s attorney should have known this allegedly missing fact. Moreover, given Luster’s claim that he had already identified Collins from the small photo, Salter’s attorney did not need the larger photo to take advantage of this information.
The alleged failure to provide these two “new facts” was not a Brady violation. The district court erred by finding that these were Brady materials and by finding a Brady violation.
III.
“[Q]ualified immunity protects all but the plainly incompetent or those who knowingly violate the law.” City of Tahlequah v. Bond , 595 U.S. 9, 12 (2021) (quotation marks and citation omitted). Salter claims that Det. Olsen’s use of the single photo was unduly suggestive, in violation of the Constitution, particularly given the severe discrepancy between Luster’s physical description of the shooters and Salter’s actual appearance. But Olsen’s testimony at Salter’s trial, quoted above, makes it clear that Olsen was proceeding on the common understanding of the law, which was that the single photo identification was lawful because Salter was not in custody, Luster knew Salter, and Luster had an independent basis for the identification. Given Olsen’s testimony, and the acceptance of that testimony by all parties at trial, including Salter’s attorney, there is no reasonable contention that Olsen was plainly incompetent or knowingly violated the law.
But the lead opinion contends that Det. Olsen, as a reasonable officer apprised of the law in 2003, must have known that a single-photo identification was unconstitutional under the circumstances of this case. If it is true that a police officer (who is not a lawyer) necessarily knows that a single-photo identification is unconstitutional, then it stands to reason that the prosecutor ( who is a lawyer ) would know that as well, and would therefore refuse to use that unconstitutional identification and instead demand more investigation before seeking an indictment. But this prosecutor did not; this prosecutor sought an indictment based almost exclusively on that single-photo identification. As I see it, either the premise is not true (i.e., the use of a single-photo identification in these circumstances was not a clearly established constitutional violation) and, therefore, the police offer (Det. Olsen) is entitled to qualified *33 immunity, or the prosecutor’s equally egregious constitutional violation is an intervening cause that cuts off the officer’s liability.
This same concern continues on for two more steps. If that premise is true—that a non- lawyer police officer (in the heat of a murder investigation) would clearly know that the single- photo identification was unconstitutional—then it stands to reason that the trial court (having time to consider opposing arguments, conduct research, and weigh its decision) would recognize the clear constitutional violation and would therefore exclude that unconstitutional identification from trial. But the trial court did not exclude the identification; the trial court determined—even on sober second thought—that the identification was lawful, and admitted it at Salter’s trial. Again, either the premise is not true, or the trial court’s constitutional mistake is an intervening cause that cuts off the officer’s liability. And taking the last step, if that premise were true, then it stands to reason that the appellate court (and the Michigan Supreme Court which denied further appeal) would certainly have known, and would have reversed the conviction due to that unconstitutional identification. Of course, that did not happen, so either the premise is not true, or the state appellate courts’ constitutional violation is an intervening cause that cuts off the officer’s liability.
Altogether, if Salter’s premise is true, then these participants—police officer, prosecutor, trial judge, appellate judges—all made the same obvious mistake, but only the officer is being held liable for it. While it is more likely that the premise is not true (i.e., it was not clearly established to a reasonable officer in 2003 that this single-photo identification was unconstitutional), even if that premise were true, there is a causation problem here that cuts off Det. Olsen’s liability.
“Like a tort plaintiff, a § 1983 plaintiff must establish both causation in fact and proximate causation.” Marvaso v. Sanchez , 971 F.3d 599, 606 (6th Cir. 2020) (quotation marks and citation omitted); Imbler v. Pachtman , 424 U.S. 409, 418 (1976) (holding that § 1983 “is to be read in harmony with general principles of tort immunities and defenses rather than in derogation of them”). “Thus, even if it is foreseeable that a defendant’s conduct will lead to the complained of harm, a defendant may be able to avoid § 1983 liability by pointing to the intervening action of a [third-party] as the proximate cause of the plaintiff’s injury.” Powers v. *34 Hamilton Cnty. Pub. Def. Comm’n , 501 F.3d 592, 610 (6th Cir. 2007). This is not novel or unusual.
In cases arising from criminal proceedings, . . . a judge often commits an intervening act. A judge may, for example, make a finding of probable cause at a preliminary hearing or issue a warrant. Such an intervening act breaks the causal chain when the judge’s action is independent from any misrepresentations, omissions, or other wrongdoing by the defendant.
Howell v. Cox , 758 F. App’x 480, 483 (6th Cir. 2018); DePiero v. City of Macedonia , 180 F.3d 770, 789 (6th Cir. 1999) (rejecting a § 1983 unreasonable seizure claim against an officer because the mayor’s court had issued the improper bench warrant). When the actions of a neutral third-party (indeed, multiple third parties) are the superseding cause of the plaintiff’s injuries, “that cuts off the otherwise foreseeable chain of causation” and “proximate cause is lacking.” Marvaso , 971 F.3d at 615 (Nalbandian, J., dissenting). That is the situation here. Even if Det. Olsen had made a clear and unmistakable constitutional error, the third parties’ (i.e., prosecutor, trial court, appellate court) failure to admit and remedy that obvious error was the proximate cause of Salter’s injury, cutting off Det. Olsen’s liability for Salter’s 15 years of wrongful imprisonment.
IV.
Because the facts material to this analysis are not contested—all of the foregoing facts are established in the record and there is no factual dispute awaiting a jury decision—we have jurisdiction to decide this interlocutory appeal, and because Det. Olsen is entitled to qualified immunity on the plain law and uncontested facts of this case, I respectfully dissent.
NOTES
[1] We recite the relevant facts in the light most favorable to Salter as we must at this stage. Berryman v. Rieger , 150 F.3d 561, 566 (6th Cir. 1998).
[2] Detective Olsen relies on Hatchett v. City of Detroit to argue that the state court’s suppression ruling precludes Salter’s claim. 495 F. App’x 567 (6th Cir. 2012). But as we have explained, “ Hatchett is an unpublished decision in which it was not clear that the criminal judgment had actually been vacated.” Peterson , 931 F.3d at 554.
[3] While Luster identified Salter as the shorter shooter (i.e., “Rob”) in a written statement taken down by Detective Olsen, he testified in court that he “never said” Salter was “Rob,” that his statement was recorded incorrectly, and that he always meant to identify Salter as the taller shooter. Trial Tr., R. 36-11, PageID 867–72.
[4] Assuming that the unlawfulness of the show-up was clearly established, the dissent posits that the prosecutor’s charging decision acts as an intervening cause that insulates Detective Olsen from liability. Dissenting Op. at 32–33. That’s not so. See Gregory , 444 F.3d at 747 (“The prosecutor’s decision to use the identification does not shield [the officer] from liability if he reasonably should have known that use of the identification would lead to a violation of Plaintiff’s right to a fair trial.”).
[1] In 2003, “the duty to disclose exculpatory evidence under Brady . . . extend[ed] only to the prosecutor, not the police.” Canter v. Cnty. of Otsego , 14 F. App’x 518, 521 n.1 (6th Cir. 2001); see Strickler v. Greene , 527 U.S. 263, 281 (1999) (“In order to comply with Brady , . . . the individual prosecutor has a duty to learn of any favorable evidence known to the others acting on the government’s behalf in this case, including the police.” (quotation marks omitted; emphasis added) (quoting Kyles v. Whitley , 514 U.S. 419, 437 (1995)). It was not until 2009, in Moldowan v. City of Warren , 578 F.3d 351 (6th Cir. 2009), that we first imposed a Brady duty on police officers, and even that was specifically limited to evidence that was “clearly exculpatory” not just “potentially useful,” id . at 383-87. The alleged Brady materials here were certainly not “clearly exculpatory.”
[2] At his deposition in 2019 Luster testified that he picked E Collins from the photo array. Putting that together with this 2003 trial testimony, one of the three photos Luster claimed to have identified was presumably Collins.
[3] Also, the record reveals that the larger photo was not withheld ; it was in the file that Det. Olsen gave to the Wayne County Prosecutor. Patricia Little, the investigator with the Conviction Integrity Unit (CIU), testified at a deposition in November 2019 that this photograph, labeled Exhibit 21 of Olsen’s Deposition, was in the Wayne County Prosecutor’s file when she received it at CIU in early 2018. See R. 36-24, PgID 994; R. 36-13, PgID 910 (photo of E Collins labeled Ex. 21 at Olsen’s Deposition). The lead opinion ignores this evidence. Instead, the lead opinion relies on inferences, assumptions, and conjecture to create the impression that the photo must not have been in the file. But this is not competing evidence for a jury to decide between. The evidence shows that the photo was in the file.