State v. Kwesi Green (080562) (Essex County and Statewide)

SYLLABUS

This syllabus is not part of the Court’s opinion. It has been prepared by the Office of the Clerk for the convenience of the reader. It has been neither reviewed nor approved by the Court. In the interest of brevity, portions of an opinion may not have been summarized.

State v. Kwesi Green (A-56/57-17) (080562) Argued January 2, 2019 -- Decided July 23, 2019 RABNER, C.J., writing for the Court.

In this case, a robbery victim identified her assailant from an extensive database of digital photos. The witness was mistakenly allowed to review the photos through a feature of the database meant to be used by law enforcement officers, not eyewitnesses. In addition, the police saved only the photo the victim ultimately selected -- an image of defendant. Beyond that, the system contained multiple photos of defendant. The Court considers what took place in light of known risks associated with eyewitness identification, as well as case law and a court rule that address how identification procedures should be conducted and preserved.

The victim, who was robbed at gunpoint while she waited for a bus, described her assailant to Newark Police Detective Donald Stabile. She said she “got a very good look” at her assailant and would be able to identify him. The victim then viewed photos through the HIDTA DataWorks PhotoManager System (HIDTA system or database). The HIDTA system for the NY/NJ region has millions of photos, including those of all adults arrested by the Newark Police Department. If someone is arrested in the relevant area more than once, the system will have multiple photos of the individual.

The HIDTA system has two modes: one for investigators and the other for witnesses. In “investigative mode,” law enforcement officials can search for a known suspect by name or other identifier. If the suspect’s identity is not known, an officer can use investigative mode to narrow the field of photos based on a witness’ description. Should the witness then select a photo and say the person in the photo resembles her assailant, an investigator can generate other similar photos to review. Individual photos can be printed, but no report of the session can be created in investigative mode. Finally, an officer can click on a photo in investigative mode to reveal a host of information about the person -- including his or her name and the date and time of arrest.

“Witness mode” is for witnesses to view digital images of mugshots. A witness can signal whether someone is the suspect or a possible suspect by telling the officer or by clicking under a photo. At the end of a session, witness mode can generate a report of the photos displayed, how long each was displayed, and whether the witness marked a

photo. The report generates a log of numbers, each of which links to a single photo. It is also possible to print individual photos in witness mode.

Detective Stabile interviewed the victim at the police station shortly after the armed robbery. He then entered various parameters in the HIDTA system in investigative mode. Without switching to witness mode, he set the witness up at a computer, explained how to scroll through photos with six on a screen at a time, and asked her to notify him if she saw her assailant or anyone who looked similar. According to the detective, the victim looked through the narrowed field of digital photos for several minutes. She then told the detective that one of the photos looked like the assailant. The detective did not know how many photos she had viewed and did not print the particular photo she flagged. Still in investigative mode, the detective narrowed the field to images similar to the one the witness selected. Within a few seconds, she identified defendant from the first page of photos she viewed next. The detective did not print the six photos from the final screen; he printed only the single photo the victim identified as her assailant. No report of the photos she viewed was generated; nor could a report have been generated from investigative mode.

Defendant was charged with first-degree robbery and weapons offenses. Defendant moved to suppress the victim’s out-of-court identification, and the trial court held an evidentiary hearing and then granted defendant’s motion. The State appealed. A majority of the Appellate Division panel found the trial court properly determined that eleven additional photos should have been preserved under Rule 3:11. However, the majority vacated the order of suppression and remanded for the trial judge to consider the full range of remedies in Rule 3:11(d). The Court granted the State’s motion for leave to appeal the order of suppression, 233 N.J. 9 (2018), and defendant’s cross-motion for leave to appeal the order of remand to reconsider the remedy, 233 N.J. 16 (2018).

HELD: Under the circumstances, the trial court properly suppressed the identification in this case. The Court proposes revisions to Rule 3:11 to offer clearer guidance on which photos officials should preserve when they use an electronic database. In addition, to guard against misidentification, the Court places on the State the obligation to show that an eyewitness was not exposed to multiple photos or viewings of the same suspect.

1. State v. Henderson reviewed concerns about the reliability of eyewitness identification and considered multiple variables that “can affect and dilute memory and lead to misidentifications.” 208 N.J. 208, 218 (2011). Among other variables, multiple viewings of a suspect can affect the reliability of an identification through “mugshot exposure” and “mugshot commitment.” Id. at 255-56. The Court in Henderson therefore observed that “law enforcement officials should attempt to shield witnesses from viewing suspects or fillers more than once.” Id. at 256. (pp. 13-16)

2. In State v. Delgado, the Court required officers to “make a written record detailing the out-of-court identification procedure, including the place where the procedure was conducted, the dialogue between the witness and the interlocutor, and the results.” 188 N.J. 48, 63 (2006). More recently, in Henderson, 208 N.J. at 252, and State v. Anthony, 237 N.J. 213, 227, 235 (2019), the Court reaffirmed those principles. (pp. 16-17)

3. The Court adopted an enhanced model jury charge in response to Henderson that includes a proposed instruction on multiple viewings and other variables. Model Jury Charge (Criminal), “Identification: In-Court and Out-of-Court Identifications” 6 (rev. July 19, 2012). Jurors are told that they “may consider whether the witness viewed the suspect multiple times during the identification process and, if so, whether that affected the reliability of the identification.” Ibid. The Court also adopted a new court rule, Rule 3:11, in response to Delgado and Henderson. However, Rule 3:11 does not address in detail what should be recorded when a witness views a database of digital photos or an electronic mug book. (pp. 17-19)

4. The Appellate Division decisions on which the State relies preceded Henderson or the date it went into effect and were decided before the adoption of Rule 3:11. Recent technological developments complicate the analysis in those opinions, and none of those cases considered Rule 3:11 or the problem of mugshot exposure. (pp. 19-22)

5. The Attorney General provided information about the use of electronic and hard copy mug books throughout the State, which the Court recounts in detail. Notably, just 3 of 479 departments use only paper mug books, in contrast to the 145 departments that use electronic or digital databases that contain photographs. (pp. 22-23)

6. Multiple views of the same person can create a risk of mugshot exposure -- the possibility that a witness will make an identification based on a memory of an earlier photo and not the original event. Henderson, 208 N.J. at 255-56. At pretrial hearings, defendants are able to explore whether an eyewitness viewed the same suspect more than once during identification procedures. Id. at 290. To guard against the risk of mugshot exposure, the Court requires the following practice going forward: When relevant, the State will have the obligation to demonstrate that an eyewitness was not exposed to multiple photos or viewings of the same suspect. If the prosecution cannot satisfy that burden, trial judges are to consider that factor when they assess whether the identification evidence can be admitted at trial. If there is a “very substantial likelihood of irreparable misidentification,” the evidence should be excluded. Id. at 289. (pp. 24-26)

7. Rule 3:11 needs to be updated. To allow for appropriate review of an out-of-court identification procedure that used a digital database or paper mug book, administrators should preserve (1) the photo of the suspect the witness selected, along with all other photos on the same screen or page, and (2) any photo that a witness says depicts a person who looks similar to the suspect, along with all other photos on that screen or page. The

Court asks the Criminal Practice Committee to revise Rule 3:11 on an expedited basis. The Court asks the Model Jury Charge Committee to amend the model charge on multiple viewings and add language about the failure to preserve an identification procedure. (pp. 26-28)

8. The Court delays the implementation of today’s ruling, aside from defendant Green, until thirty days from the date the Court approves revisions to Rule 3:11, at which time this decision will apply prospectively. (p. 28)

9. When the record of an identification “is lacking in important details,” and it was feasible to preserve them, Rule 3:11(d) affords a judge discretion, consistent with appropriate case law, to bar the evidence, redact part of it, and/or “fashion an appropriate jury charge” if the evidence is admitted. Under the circumstances, the Court cannot find that the trial judge abused his discretion when he suppressed the identification. The Court does not suggest that any time a full record of an identification is not preserved, the evidence must be excluded. Indeed, suppression should be the remedy of last resort, and judges should explain why other remedies in Rule 3:11(d) are not adequate before barring identification evidence. It is the confluence of factors in this appeal -- the witness’s use of investigative mode during the identification process, the failure to preserve all but one photo, and defendant’s history of multiple recent arrests, which would have been captured in the HIDTA system the witness viewed -- that casts doubt on the reliability of the identification process and supports the trial court’s conclusion. (pp. 28-30)

The judgment of the Appellate Division is AFFIRMED AS MODIFIED.

JUSTICE SOLOMON, concurring in part and dissenting in part, concurs in the majority’s ruling but not in its remedy. Justice Solomon explains that the trial court discussed only whether the evidence should be suppressed and, in doing so, did not measure the evidence in the record against the prevailing legal standard: whether defendant had shown a “very substantial likelihood of irreparable misidentification” under the circumstances. See Henderson, 208 N.J. at 289. Because a trial court must make that determination before deciding whether the high threshold for suppression has been vaulted, Justice Solomon would affirm the Appellate Division and remand to the trial court to “probe what happened during the identification process” and meaningfully assess the identification’s reliability. Anthony, 237 N.J. at 239.

JUSTICES LaVECCHIA, ALBIN, PATTERSON, FERNANDEZ-VINA, and TIMPONE join in CHIEF JUSTICE RABNER’s opinion. JUSTICE SOLOMON filed a separate opinion, concurring in part and dissenting in part.

SUPREME COURT OF NEW JERSEY A-56/57 September Term 2017 080562

State of New Jersey,

Plaintiff-Appellant/Cross-Respondent,

v.

Kwesi Green,

Defendant-Respondent/Cross-Appellant.

On appeal from the Superior Court, Appellate Division.

Argued Decided January 2, 2019 July 23, 2019

Lucille M. Rosano, Special Deputy Attorney General/Assistant Prosecutor, argued the cause for appellant/cross-respondent (Theodore N. Stephens, II, Acting Essex County Prosecutor, attorney; Lucille M. Rosano, of counsel and on the briefs).

Peter T. Blum, Assistant Deputy Public Defender, argued the cause for respondent/cross-appellant (Joseph E. Krakora, Public Defender, attorney; Peter T. Blum, of counsel and on the briefs).

Lauren Bonfiglio, Deputy Attorney General, argued the cause for amicus curiae Attorney General of New Jersey (Gurbir S. Grewal, Attorney General, attorney; Lauren Bonfiglio, of counsel and on the brief).

Lawrence S. Lustberg argued the cause for amici curiae Innocence Project, American Civil Liberties Union of New Jersey, and Innocence Network (Gibbons and American Civil Liberties Union of New Jersey Foundation, attorneys; Lawrence S. Lustberg, Farbod K. Faraji, and Alexander Shalom, on the brief).

CHIEF JUSTICE RABNER delivered the opinion of the Court.

In this case, a robbery victim identified her assailant from an extensive database of digital photos. To assess the reliability of the identification process requires an understanding of modern-day digital databases.

In some respects, they are today’s equivalent of a paper mugshot book. In other ways, digital systems are far superior, thanks to advances in technology. The system used here, for example, allows officers to pare down a large field of photos to match a witness’s physical description of a suspect. When an eyewitness selects a photo that looks similar to the culprit, the system can further narrow the field to display only other similar images. Officers can also print copies of photos and generate a report of what a witness viewed.

In this appeal, the witness was mistakenly allowed to review digital photos through a feature of the database meant to be used by law enforcement officers, not eyewitnesses. In addition, the police saved only the photo the victim ultimately selected -- an image of defendant. Beyond that, the system contained multiple photos of defendant because of his recent prior arrests, *3 which raises concerns about mugshot exposure and its effect on the reliability of identifications.

We consider what took place in light of known risks associated with eyewitness identification, as well as case law and a court rule that address how identification procedures should be conducted and preserved. We also propose revisions to Rule 3:11 to offer clearer guidance on which photos officials should preserve when they use an electronic database to identify a suspect. In addition, to guard against misidentification, we place on the State the obligation to show that an eyewitness was not exposed to multiple photos or viewings of the same suspect.

Under the circumstances, we find that the trial court properly suppressed the identification in this case. We therefore affirm and modify the judgment of the Appellate Division majority, which largely upheld the trial court.

I.

To recite the facts, we draw heavily on the testimony at the suppression hearing. On February 11, 2014, a woman was robbed at gunpoint while she waited at a bus stop in Newark. A man approached her, pointed a black handgun at her chest, demanded her pocketbook, and asked several questions about other items she was carrying. The woman handed her bag to the man, who walked away.

Soon after, the victim described her assailant to Newark Police Detective Donald Stabile and another officer at the police station. She said the man was about 5’7” tall, weighed between 130 and 150 pounds, had dark skin, a round face, short hair, and no facial hair, and was in his early twenties. The victim said she “got a very good look” at her assailant and would be able to identify him.

Based on her description, Detective Stabile input certain criteria into a digital database of photos, from which the victim ultimately identified defendant Kwesi Green. Because the identification is central to this appeal, we describe in detail both the database and the process the police followed.

A.

The victim viewed photos through the HIDTA DataWorks PhotoManager System (HIDTA system or database). HIDTA -- an abbreviation for High Intensity Drug Trafficking Area -- is a federally funded grant program that provides assistance to state and local law enforcement agencies to combat drug trafficking in a coordinated manner. The Newark Police Department participates in the New York/New Jersey HIDTA region, which covers seventeen counties in New York and northern New Jersey, including Essex County.

The HIDTA system for the NY/NJ region has millions of digital photos, *5 including photos of all adults arrested by the Newark Police Department. 1 When officers who participate in HIDTA process a defendant, they enter the individual’s arrest photo in the PhotoManager System. If someone is arrested in the relevant area more than once, the system will have multiple photos of the individual -- one for each arrest.

The HIDTA system has two modes: one for investigators and the other for witnesses. “Investigative mode” -- as its name suggests -- is designed for law enforcement officials. Investigators can search for a known suspect by name, date of birth, or some other identifier; if the suspect’s identity is not known, an officer can use investigative mode to narrow the field of photos for a witness to view. Specifically, if a witness describes an unknown person’s age, race, height, weight, hair color, facial features, or other pedigree information, an officer can tailor a search to match those parameters. Should

1 Detective Stabile, who testified he had used Newark’s HIDTA system hundreds of times, said the database contained photos only of individuals arrested by the Newark Police Department. A second witness, Robert Vitale, testified the system had photos of individuals arrested in all seventeen counties. At the time of the hearing, Mr. Vitale was employed by DataWorks as the senior systems engineer for the New York City Police Department’s PhotoManager system. He supervised the design, installation, and maintenance of that system. Mr. Vitale testified that he was familiar with the Newark HIDTA system and had spoken with DataWorks engineers responsible for it, but he had not worked on a Newark terminal. To the extent there is a discrepancy in the record about which photos are included in the Newark system, the issue is not material to this appeal.

the witness then select a photo and say the person in the photo resembles her assailant, an investigator can highlight that photo and generate other similar photos to review. Individual photos displayed or selected in investigative mode can be printed, but no report of the session can be created.

Using a feature for similar images or photos, an officer can also generate a photo array -- a group of photos that look like a certain image -- in investigative mode. As the array is created, particular photos will reappear until the officer replaces them with other images. The final array can be saved.

In addition, an officer can click on a photo in investigative mode to reveal a host of information about the person -- including his or her name and the date and time of arrest.

“Witness mode” is for witnesses to view digital images of mugshots. It is similar to using an old-fashioned, paper mug book. A witness can signal whether someone is the suspect or a possible suspect by telling the officer or, as Mr. Vitale testified, by clicking under a photo. At the end of a session, witness mode can generate a report of the photos displayed, how long each was displayed, and whether the witness marked a photo, according to Mr. Vitale. The report generates a log of identifier numbers, each of which links to a single photo. It is also possible to print individual photos in witness mode. *7 Mr. Vitale testified that the Newark Police Department had the ability to use witness mode in February 2014 -- when the robbery took place.

The two modes work in tandem. For example, an officer can first narrow the universe of photos in investigative mode, based on a witness’s description. Next, the officer can create a session in witness mode for the witness to examine the images. If the witness identifies a suspect, the session is complete, and a report can be generated. If the witness instead identifies a photo that looks similar to the suspect, the officer can close out the session, switch back to investigative mode, and search for additional similar photos to narrow the field further. The witness can then examine the more limited group of photos as part of a new session in witness mode. Each session from witness mode can be preserved, and officers can print individual photos from either mode.

B.

As noted earlier, Detective Stabile interviewed the victim at the police station shortly after the armed robbery. He then entered various parameters in the HIDTA system based on the description of the assailant’s height, weight, age, skin color, and hair. The search was done in investigative mode. Without switching to witness mode, the detective set the witness up at a computer monitor, explained how to scroll through pages of photos, with six on a screen *8 at a time, and asked her to notify him immediately if she saw her assailant or anyone who looked similar.

According to the detective, the victim looked through the narrowed field of digital photos for several minutes. She then told the detective that one of the photos looked like the assailant. The detective did not know how many photos she had viewed and did not print the particular photo she flagged. Still in investigative mode, the detective narrowed the field to images similar to the one the witness selected. Within a few seconds, she identified defendant from the first page of photos she viewed next. The detective could not recall if the photo the witness had said looked similar to her assailant was repeated among the six photos from which she made her identification.

The detective did not print the six photos from the final screen; he printed only the single photo the victim identified as her assailant. No report of the photos she viewed was generated; nor could a report have been generated from investigative mode.

C.

The grand jury returned an indictment that charged defendant with first-degree robbery, N.J.S.A. 2C:15-1; second-degree unlawful possession of a handgun, N.J.S.A. 2C:39-5(b); and second-degree possession of a weapon for an unlawful purpose, N.J.S.A. 2C:39-4(a). Defendant moved to suppress the *9 victim’s out-of-court identification, and the trial court held an evidentiary hearing at which Detective Stabile and Mr. Vitale testified. Afterward, the court suppressed the identification.

The court found the police had conducted an identification procedure within the meaning of Rule 3:11. Under the Rule and relevant case law, the court concluded that officers should have preserved the image the witness said looked like the suspect, the five other images displayed on that screen, a s well as the final six images that included defendant’s photo -- a total of eleven photos beyond the one the officer preserved. The court noted it was feasible to print all eleven images.

In light of the record developed at the hearing, the trial judge distinguished State v. Joseph, 426 N.J. Super. 204 (App. Div. 2012). Joseph involved a witness who reviewed photos in the HIDTA database. Id. at 213. The Appellate Division in that case concluded that law enforcement’s failure to preserve all of the photos the victim viewed was “not fatal to the admission of an out-of-court identification.” Id. at 223.

Joseph, in turn, relied on cases in which witnesses viewed random groups of photos in a paper mug book and on a computer screen. Ibid. (discussing State v. Janowski, 375 N.J. Super. 1, 6-8 (App. Div. 2005); State v. Ruffin, 371 N.J. Super. 371, 395 (App. Div. 2004)). By contrast, the trial *10 court here found, the “application of the HIDTA system [is] not random.” As the court explained, the detective entered search parameters from the victim’s general description to generate a series of images, and later narrowed the universe of photos when the witness selected an image that looked similar to her assailant. Once the parameters were entered, “[i]t was no longer random.” Moreover, the court observed, “[a]ny suggestion” the process was “random end[ed] at the point where the witness said that looks like him and the officer hit [the] similar” function.

Because of the officer’s failure to preserve the other eleven photos, the court suppressed the out-of-court identification. The trial court’s ruling did not extend to any in-court identification at trial. 2

The State appealed. A majority of the Appellate Division panel concluded that “the plain language” of Rule 3:11 “requires that when digital photos are shown during an identification procedure, a record of the photos used must be made.” The majority observed that its interpretation was consistent with the history of the Rule, the Court’s “longstanding policy” to

2 The Innocence Project and related amici voice concerns about the suggestive nature of in-court identifications that follow an earlier identification procedure. Because the issue has not been raised by the parties, we do not address it at this time. See State v. O’Driscoll, 215 N.J. 461, 479 (2013); Bethlehem Twp. Bd. of Educ. v. Bethlehem Twp. Educ. Ass’n, 91 N.J. 38, 48- 49 (1982).

provide “broad discovery” to criminal defendants, and “concerns about identification procedures” discussed in State v. Delgado, 188 N.J. 48 (2006), and State v. Henderson, 208 N.J. 208 (2011). Among other things, the majority pointed to the risk of mugshot exposure from multiple views of photos of the same person.

The majority found the trial court properly determined that eleven additional photos should have been preserved under Rule 3:11. However, the majority vacated the order of suppression and remanded for the trial judge to consider the full range of remedies in Rule 3:11(d).

The dissenting Appellate Division judge relied on Ruffin, Janowski, and Joseph to reach a contrary result. It also observed that “there is no police suggestiveness to be counteracted” when victims view mug books for unknown suspects. In the dissent’s view, Rule 3:11 does not “address the use of . . . computerized mug books to search for an unknown perpetrator”; nor does the Rule’s history. To require the preservation of such photos, the dissent added, would “needlessly burden” law enforcement.

We granted the State’s motion for leave to appeal the order of suppression, 233 N.J. 9 (2018), and defendant’s cross-motion for leave to appeal the order of remand to reconsider the remedy, 233 N.J. 16 (2018). We also granted leave to appear as amici curiae to the Attorney General and to the *12 Innocence Project, the American Civil Liberties Union of New Jersey, and the Innocence Network (collectively, the Innocence Project), which filed a joint brief.

After oral argument, we asked for supplemental briefing about the use of electronic and hard copy mug books by law enforcement throughout the State. We also asked for details about defendant’s prior arrest record.

II.

The State argues that Rule 3:11 does not apply when the HIDTA system is used to investigate an unknown suspect. As a result, the State submits, the Rule does not require the preservation of the additional photographs the trial court ordered. In the alternative, if the Rule was violated, the State contends that a properly tailored jury instruction, not suppression, is the appropriate remedy.

The Attorney General echoes those arguments. He contends that Rule 3:11 does not apply to the review of random mugshots to try to locate a suspect. Law enforcement must preserve only the photo the witness identified, the Attorney General argues.

Defendant submits that case law, Rule 3:11, and “fundamental notions of justice” required law enforcement to preserve the photos used in the HIDTA system. In addition, defendant argues that the trial court properly suppressed *13 the identification because the use of the HIDTA system was “prone to create a mugshot exposure problem” and the State could not reconstruct “the lost trail of photos.”

The Innocence Project similarly contends that Rule 3:11 applies to out-of-court identifications made with the use of the HIDTA database, even when the police do not have a known suspect. Amicus also maintains that social science principles about the reliability of eyewitness identification evidence apply with equal force to law enforcement’s use of the HIDTA database. Finally, the Innocence Project submits that suppression is appropriate here.

III.

A.

This appeal raises questions about identification evidence in the context of electronic mug books. Henderson reviewed concerns about the reliability of eyewitness identification evidence more broadly. The Court discussed how “memory is malleable” and observed that “eyewitness misidentification is the leading cause of wrongful convictions” nationwide. Henderson, 208 N.J. at 218.

Henderson considered multiple variables that “can affect and dilute memory and lead to misidentifications.” Ibid. The factors include “system variables” that law enforcement can control, like the use of a lineup or a mug *14 book, and “estimator variables” that are outside the control of the legal system, like lighting and memory decay. See ibid.

The Court ultimately determined that the then-existing standard to evaluate eyewitness identifications -- derived from Manson v. Brathwaite, 432 U.S. 98 (1977), and State v. Madison, 109 N.J. 223 (1988) -- did “not offer an adequate measure for reliability or sufficiently deter inappropriate police conduct.” Henderson, 208 N.J. at 218. As a result, the Court “revised the legal framework for the admission of eyewitness identification evidence .” State v. Anthony, 237 N.J. 213, 226 (2019). Today, if defendants can present some evidence of suggestiveness related to a system variable, they are entitled to a hearing to explore all relevant factors and challenge the admission of the identification evidence. Ibid. (citing Henderson, 208 N.J. at 288-93).

Among other variables, the Court has recognized that multiple viewings of a suspect can affect the reliability of an identification. Henderson, 208 N.J. at 255. With “successive views of the same person,” it can be “difficult to know whether the later identification stems from a memory of the original event or a memory of the earlier identification procedure.” Ibid.

The problem can surface when a witness uses mugshot books to make an identification. If a book contains multiple arrest photos of the same person, there is a risk of “mugshot exposure.” See ibid. That happens, for example, *15 “when a witness initially views a set of photos and makes no identification, but then selects someone -- who had been depicted in the earlier photos -- at a later identification procedure.” Ibid. Multiple studies have revealed that “although 15% of witnesses mistakenly identified an innocent person viewed in a lineup for the first time, that percentage increased to 37% if the witness had seen the innocent person in a prior mugshot.” Id. at 255-56 (citing Kenneth A. Deffenbacher et al., Mugshot Exposure Effects: Retroactive Interference, Mugshot Commitment, Source Confusion, and Unconscious Transference, 30 L. & Hum. Behav. 287, 299 (2006)).

“Mugshot commitment” is a related concern. It “occurs when a witness identifies a photo that is then included in a later lineup procedure. Studies have shown that once witnesses identify an innocent person from a mugshot, ‘a significant number’ then ‘reaffirm[] their false identification’ in a later lineup -- even if the actual target is present.” Id. at 256 (quoting Gunter Koehnken et al., Forensic Applications of Line-Up Research, in Psychological Issues in Eyewitness Identification 205, 219 (Siegfried L. Sporer et al. eds., 1996)).

Because mugshot exposure and commitment can affect the reliability of an identification and heighten the risk of misidentification, the Court in Henderson observed that “law enforcement officials should attempt to shield witnesses from viewing suspects or fillers more than once.” Ibid. Henderson, *16 however, did not specifically address electronic mug books or digital databases of arrest photos.

Henderson is consistent with guidance that the U.S. Department of Justice issued years earlier. In its 1999 guide for law enforcement on eyewitness evidence, the Justice Department discussed how investigators should prepare mug books. Nat’l Inst. of Justice, U.S. Dep’t of Justice, Eyewitness Evidence: A Guide for Law Enforcement 17-18 (1999). The Department stressed that mug books should be composed “in such a manner that individual photos are not suggestive.” Id. at 17. To that end, the Department stated that “preparer[s] should . . . [e]nsure that only one photo of each individual is in the mug book.” Ibid.

B.

In Anthony, we recently reviewed another important topic: the need to make a record of identification procedures. Anthony traced the issue through a line of cases that dates back decades. In 1972, in State v. Earle, this Court explained that

enforcement authorities should . . . make a complete record of an identification procedure if it is feasible to do so . . . . The identity of persons participating in a [live] lineup should be recorded, and a picture should be taken if it can be. If the identification is made or attempted on the basis of photographs, a record should be made of the photographs exhibited. We do not say

a failure hereafter to follow such procedures will itself invalidate an identification, but such an omission, if not explained, should be weighed in deciding upon the probative value of the identification, out-of-court and in-court.

[60 N.J. 550, 552 (1972).]

In Delgado, the Court required law enforcement officers to “make a written record detailing the out-of-court identification procedure, including the place where the procedure was conducted, the dialogue between the witness and the interlocutor, and the results.” 188 N.J. at 63. More recently, in Henderson, 208 N.J. at 252, and Anthony, 237 N.J. at 227, 235, the Court reaffirmed the principles outlined in Delgado.

Those earlier decisions, however, did not focus specifically on what must be preserved when a witness makes an identification from a paper or electronic mug book.

C.

The Court adopted an enhanced model jury charge in response to Henderson. It includes a proposed instruction on multiple viewings and other variables. The instruction informs the jury that multiple viewings of the same person increase the risk of mistaken identification. Model Jury Charge (Criminal), “Identification: In-Court and Out-of-Court Identifications” 6 (rev. July 19, 2012). As a result, jurors are told that they “may consider whether the *18 witness viewed the suspect multiple times during the identification process and, if so, whether that affected the reliability of the identification.” Ibid.

The Court also adopted a new court rule in response to Delgado and Henderson. See R. 3:11; see also Anthony, 237 N.J. at 229-30 (tracing the Rule’s history). We reviewed Rule 3:11 at length in Anthony and focused on law enforcement’s obligation to record the dialogue between a witness and an officer. 237 N.J. at 230-34.

The focus in this case is on how to record the identification process when a witness selects a suspect from a series of photos. Rule 3:11 addresses the topic in the following two sections:

(a) Recordation. An out-of-court identification resulting from a photo array, live lineup, or showup identification procedure conducted by a law enforcement officer shall not be admissible unless a record of the identification procedure is made.

....

(c) Contents. The record of an out-of-court identification procedure is to include details of what occurred at the out-of-court identification, including the following:

....

(5) if a photo lineup, the photographic array, mug books or digital photographs used . . . .

As is apparent from its text, the Rule does not address in detail what should be recorded when a witness views a database of digital photos or an electronic mug book.

Rule 3:11 also sets forth a number of remedies when the record is deficient:

If the record that is prepared is lacking in important details as to what occurred at the out-of-court identification procedure, and if it was feasible to obtain and preserve those details, the court may, in its sound discretion and consistent with appropriate case law, declare the identification inadmissible, redact portions of the identification testimony, and/or fashion an appropriate jury charge to be used in evaluating the reliability of the identification.

[R. 3:11(d).]

D.

The State relies on several Appellate Division decisions that preceded Henderson or the date it went into effect, and were decided before the adoption of Rule 3:11. The rulings addressed the use of paper or electronic mug books to search for an unknown suspect and, in each case, the Appellate Division concluded the police needed to preserve only the photo of the culprit the victim selected.

1.

In State v. Ruffin, a burglary victim looked through two-and-a-half *20 loose-leaf binders of arrest photos of black males -- more than 400 photos altogether. 371 N.J. Super. at 378-79. The photos were arranged randomly; “[t]here was no separation based on height, weight, hair style, facial hair or complexion.” Id. at 379, 395. Because the police did not preserve the binders, the trial court excluded the victim’s out-of-court identification. Id. at 380.

The Appellate Division reversed. The court observed that the process “was investigatory, not confirmatory,” and that the presentation of the random photographs was “neutral.” Id. at 395. The Appellate Division added that it would be “cumbersome” to require the police to preserve all of the photos the witness viewed, which “would also place an unnecessary burden on” the use of a “proper law enforcement tool for no justifiable purpose.” Ibid.

In State v. Janowski, a robbery victim identified an unknown assailant from 36 to 60 arrest photos displayed on a computer screen. 375 N.J. Super. at 4-5. The photos depicted white men “within the age range the victim provided.” Id. at 5.

The Appellate Division rejected the trial court’s finding that the photos constituted an array, which must be preserved in full. Id. at 5, 7. Instead, the Appellate Division considered each screen “the equivalent of a page of a mug shot book.” Id. at 7. Relying on Ruffin, the court held that the State’s failure to retain all of the “randomly selected photographs, kept for the purpose of *21 investigation, not confirmation,” and shown to the victim to try to find a suspect, was “not fatal to the admission of the out-of-court identification.” Id. at 8-9.

More recently, in State v. Joseph, three victims robbed at gunpoint in Newark identified an unknown assailant from the same HIDTA database used in this case. 426 N.J. Super. at 212-15. The Court applied the same reasoning used in Ruffin and Janowski and concluded the police needed to preserve only the photo the victims selected. Id. at 223.

2.

Recent developments complicate the analysis in those opinions. First, developments in technology muddy the clear-cut divide earlier decisions relied on between an examination of random photos in a mug book, which historically did not have to be preserved, and the use of potentially suggestive photo arrays, which plainly did.

None of the decisions, even Joseph, delve into the features of the HIDTA PhotoManager System; presumably, the record in Joseph did not include the details before the Court today. Yet testimony at the evidentiary hearing reveals that, unlike a paper mug book, the HIDTA system offers more than a random, digital assortment of photos.

In this case, a detective first entered information about the suspect’s *22 race, age, weight, and appearance into the database. The victim then selected a photo of someone that she believed looked like the robber. Based on her feedback, the detective further narrowed the search results to look for other similar images. By that point at least, the search was no longer random. Nor was it a photo array with a known suspect. The identification process here, made possible by a more sophisticated computer program, lies somewhere along the continuum between random mugshots in a book and potentially suggestive photo arrays with a known suspect.

Second, as noted above, all three Appellate Division cases preceded developments in the case law and court rules. Ruffin and Janowski predated Henderson, and Joseph properly did not apply Henderson retroactively. See Joseph, 426 N.J. Super. at 225 n.5. All three decisions preceded Rule 3:11. As a result, none of the cases considered the Rule or the problem of mugshot exposure, which the HIDTA system presents. See Henderson, 208 N.J. at 255-56.

IV.

After oral argument, the Court asked the Attorney General to provide information about the use of electronic and hard copy mug books throughout the State. The Attorney General surveyed 479 “police departments,” including every municipal police department, the Delaware River Port Authority, Morris *23 County Park Police, New Jersey Human Services, New Jersey Transit, Palisades Interstate Parkway, and campus police. The survey revealed the following:

 6 of 479 police departments use hard copy or paper mug books as

part of a department’s identification procedure. Three of those

departments use only hard copy or paper mug books.

 145 of 479 police departments use electronic or digital databases

that contain photographs. The rest do not use that type of

identification procedure.

 22 of the 145 police departments that use electronic or digital

databases use the HIDTA Digital PhotoManager system. The

other departments use any of more than two dozen different

systems.

 Neither the State Police nor the Division of Criminal Justice use

paper mug books or digital databases as part of their identification

procedures.

 No County Prosecutor’s office uses paper mug books; ten use

various digital databases, including the HIDTA Digital

PhotoManager System.

V.

A.

Because of the nature of the HIDTA system, we return to the issue of multiple viewings of the same person during an identification procedure. Mug books are made up of arrest photos, and if an individual has been arrested more than once, multiple photos of the person will appear in both a paper mug book and a digital database.3

As noted earlier, multiple views of the same person can create a risk of mugshot exposure -- the possibility that a witness will make an identification based on a memory of an earlier photo and not the original event. Henderson, 208 N.J. at 255-56. Research reviewed in Henderson showed that mistaken identifications increased from 15 to 37% when a witness had seen a photo of an innocent person in a prior mugshot. Ibid. (citing Deffenbacher, 30 L. & Hum. Behav. at 299). In light of those results, it is not surprising that the Department of Justice issued guidance to law enforcement to “[e]nsure that only one photo of each individual is in the mug book.” Nat’l Inst. of Justice at 17.

3 In the HIDTA system’s investigative mode, if a witness selects a photo that looks like a suspect, and an investigator narrows the field to similar images, it is unclear from the record whether the look-alike photo will reappear. In this case, the detective did not recall whether the photo that looked similar to the suspect appeared twice.

At pretrial hearings, defendants are now able to explore whether an eyewitness viewed the same suspect more than once during identification procedures. Henderson, 208 N.J. at 290. To guard against the risk of mugshot exposure, we exercise our supervisory powers under Article VI, Section 2, Paragraph 3 of the State Constitution to require the following practice going forward: When relevant, the State will have the obligation to demonstrate that an eyewitness was not exposed to multiple photos or viewings of the same suspect. See id. at 254, 270-71 (exercising supervisory powers relating to the identification process); Delgado, 188 N.J. at 63 (same). If the prosecution cannot satisfy that burden, trial judges are to consider that factor when they assess whether the identification evidence can be admitted at trial. If there is a “very substantial likelihood of irreparable misidentification,” the evidence should be excluded. Henderson, 208 N.J. at 289 (citing Manson, 432 U.S. at 116).

We recognize the above approach will create a practical challenge for the few departments that rely solely on paper or hard copy mug books. That said, only 3 of nearly 500 departments statewide still use that method exclusively. The same obligation will apply to the 145 departments that use electronic databases. In that regard, the record in this case shows how sophisticated the HIDTA PhotoManager System is. It can narrow a field of *26 photos based on a suspect’s age, height, weight, skin color, hair, and facial hair. A defendant’s name and an identifier number are attached to each photo as well. It stands to reason, then, that technical solutions can be found to eliminate multiple photos of the same person in commercially available electronic databases. 4

Digital databases already include measures to prevent witnesses from gaining access to extraneous, possibly prejudicial information about a person’s arrest history. With the benefit of proper training, law enforcement officers should use features like the HIDTA PhotoManager System’s witness mode when they ask witnesses to view digital or electronic mug books.

B.

We note as well that Rule 3:11 needs to be updated. The parties present thoughtful views about the text of the current Rule. Those arguments highlight the need for greater clarity in this area. We rely on our rulemaking authority, N.J. Const. art. VI, § 2, ¶ 3, to offer clearer guidance about the type of evidence law enforcement should preserve when a witness identifies a suspect

4 DataWorks’s current website says its “system can be configured to not allow multiple images of an individual to a lineup.” Digital PhotoManager, DataWorks Plus, http://www.dataworksplus.com/dpm.html (last visited June 27, 2019). The site does not say how that is accomplished or how long the feature has been available. Testimony at the suppression hearing did reveal that a log of unique identifier numbers can be generated in witness mode, and that each number is linked to an individual photo.

from a digital or electronic database.5 For that reason, we do not address at length certain arguments raised by the parties about the meaning of the current Rule. See Anthony, 237 N.J. at 230 & n.2.

To allow for appropriate review of an out-of-court identification procedure that used a digital database or paper mug book, administrators should preserve (1) the photo of the suspect the witness selected, along with all other photos on the same screen or page, and (2) any photo that a witness says depicts a person who looks similar to the suspect, along with all other photos on that screen or page. That requirement will establish a record of the photos viewed at pivotal moments: when the witness meaningfully narrows the field of images and ultimately makes a final selection. It will also allow all parties and the court to assess with care the nature of the identification process and any suggestive aspects of the process. Aside from the risk of multiple viewings addressed above, persuasive reasons have not been presented to require that every photo reviewed during the entire identification procedure be preserved.

We ask the Criminal Practice Committee to revise Rule 3:11 consistent with the above concepts. We also ask the Model Jury Charge Committee to

5 We use the terms “digital” and “electronic” interchangeably in this opinion. The opinion and any revisions to Rule 3:11 should be interpreted to encompass any similar medium as well.

amend the model charge on multiple viewings and add language about the failure to preserve an identification procedure. See id. at 235.

We recognize it will take time for the Committee to propose changes to the Rule. Likewise, the Attorney General will need time to disseminate guidance to the law enforcement community about the proper use of digital databases, including their different modes, and the specific photos that must be preserved. Aside from defendant Green, we therefore delay the implementation of today’s ruling until thirty days from the date the Court approves revisions to Rule 3:11, at which time this decision will apply prospectively. See Henderson, 208 N.J. at 302 (applying ruling prospectively because of the effect on the administration of justice). We respectfully ask the Committee to propose revisions on an expedited basis.

VI.

We next consider the remedy in this case. The detective who administered the identification procedure preserved only one photo -- the one the victim identified as her assailant. The trial court concluded the detective should have preserved the eleven additional photos described above.

In our judgment, the manner in which the detective conducted the identification procedure is also problematic. It was a mistake for the victim to be allowed to view photos of possible suspects through the HIDTA system’s *29 investigative mode. As noted above, that mode is designed for investigators to narrow a large field of photos based on a witness’s description and feedback.

Witnesses should not examine photos in investigative mode for a number of reasons. That mode allows eyewitnesses to access details that witnesses should not see, like a defendant’s name and the date and place of arrest. It repeats individual photos when an array is created. And it cannot generate a report of what an eyewitness viewed. By contrast, witness mode limits access to potentially suggestive information; allows witnesses to examine suitable and narrowed fields of photos; and enables administrators to keep a record of the session.

Finally, we note the real possibility of mugshot exposure in this case. The armed robbery took place in Newark on February 11, 2014. In the four months leading up to that date, defendant was arrested twice in Newark as an adult. Photos from both arrests would have been part of Newark’s HIDTA system, which the victim viewed. In addition, the detective could not recall whether the photo the witness said looked like the culprit appeared a second time on the screen that included defendant Green’s photo.

When the record of an identification “is lacking in important details,” and it was feasible to preserve them, Rule 3:11(d) affords a judge discretion, consistent with appropriate case law, to bar the evidence, redact part of it, *30 and/or “fashion an appropriate jury charge” if the evidence is admitted. Under the circumstances, we cannot find that the trial judge abused his discretion or ruled in a manner that was inconsistent with appropriate case law when he suppressed the identification.

We do not suggest that any time a full record of an identification is not preserved, the evidence must be excluded. To be clear, we do not adopt a per se rule to that effect. See Anthony, 237 N.J. at 239; Henderson, 208 N.J. at 303. Indeed, suppression should be the remedy of last resort, and judges should explain why other remedies in Rule 3:11(d) are not adequate before barring identification evidence. It is the confluence of factors in this appeal -- the witness’s use of investigative mode during the identification process, the failure to preserve all but one photo, and defendant’s history of multiple recent arrests in Newark, which would have been captured in the HIDTA system the witness viewed -- that casts doubt on the reliability of the identification process and supports the trial court’s conclusion.

VII.

For the reasons outlined above, the judgment of the Appellate Division is affirmed as modified.

JUSTICES LaVECCHIA, ALBIN, PATTERSON, FERNANDEZ-VINA, and TIMPONE join in CHIEF JUSTICE RABNER’s opinion. JUSTICE SOLOMON filed a separate opinion, concurring in part and dissenting in part.

State of New Jersey,

Plaintiff-Appellant/Cross-Respondent,

v.

Kwesi Green,

Defendant-Respondent/Cross-Appellant.

JUSTICE SOLOMON, concurring in part and dissenting in part.

I agree with the majority that there are several reasons why the detective’s identification procedure was problematic. I also agree that the identification record was “lacking in important details.” R. 3:11(d). However, I cannot agree that the trial judge complied with Rule 3:11 and applicable case law in suppressing the identification. Because this Court’s jurisprudence makes clear that a remand is appropriate here, and not suppression of the identification, I concur in the majority’s ruling but not in its remedy.

When the record of an identification “is lacking in important details,” and it was feasible to preserve them, Rule 3:11(d) affords a judge discretion, consistent with appropriate case law, to bar the evidence, redact part of it, “and/or fashion an appropriate jury charge” if the evidence is admitted. With that in mind, this Court has consistently rejected the invitation to “create[] bright-line rules that call for the ‘suppression of reliable evidence any time a *2 law enforcement officer makes a mistake.’” State v. Anthony, 237 N.J. 213, 239 (2019) (quoting State v. Henderson, 208 N.J. 208, 303 (2011)). We have done so for good reason; the threshold for suppression has always been high and remains so today. See ibid.; see also Henderson, 208 N.J. at 303. Unless a defendant “demonstrate[s] a very substantial likelihood of irreparable misidentification,” Henderson, 208 N.J. at 289, it must be left “for the jury to decide whether to credit a witness’ account, with the benefit of the augmented model jury charge,” Anthony, 237 N.J. at 239. As such, remand hearings can serve an important purpose -- to “probe what happened during the identification process [] and end with evidence being excluded if it is unreliable, and admitted otherwise.” Ibid.

Here, the trial court did not address the full range of possible remedies; it discussed only whether the evidence should be suppressed. In doing so, the court also did not measure the evidence in the record against the prevailing legal standard: whether defendant Kwesi Green had shown a “very substantial likelihood of irreparable misidentification” under the circumstances. See Henderson, 208 N.J. at 289. Because a trial court must make that determination before deciding whether the high threshold for suppression has been vaulted, I would affirm the Appellate Division and remand to the trial court to “probe what happened during the identification process” and *3 meaningfully assess the identification’s reliability. Anthony, 237 N.J. at 239. In my view, suppression of the identification, particularly in light of the trial court’s limited findings here, may preclude the jury from performing its function.