The New Jersey Supreme Court has issued a significant ruling reshaping how defendants can access information about facial recognition technology used against them during criminal investigations, establishing that such discovery requests must be evaluated on a case by case basis rather than through a single, uniform checklist applied automatically to every case involving the technology. The decision, handed down in State v. Tybear Miles on June 24, 2026, clarifies a legal question that had grown increasingly urgent as facial recognition tools become a more common feature of law enforcement investigations across New Jersey and the rest of the country.
The case itself grew out of a fatal shooting in Jersey City that led a grand jury to charge defendant Tybear Miles with first degree murder along with related weapons offenses. During their investigation, police worked with a confidential informant who reviewed surveillance footage captured near the scene, identifying two individuals by their nicknames and providing Instagram usernames connected to each of them. Investigators then ran a facial recognition search against the Instagram profile photograph associated with one of those individuals, a search that returned Miles as a match alongside several other possible candidates. Police subsequently interviewed Miles’s sister, his former girlfriend, and two additional individuals, each of whom identified him after viewing video and still images pulled from separate nearby surveillance footage. Notably, that surveillance footage showed multiple men in the area, none of it actually captured the shooting itself, and no witness ever identified Miles as the person who fired the fatal shots.
Given the central role facial recognition technology played in connecting Miles to the investigation in the first place, his defense team moved to compel discovery related to that technology, relying on the precedent set in State v. Arteaga, a 2023 Appellate Division decision that had granted a defendant in a similar situation access to thirteen specific categories of information regarding the facial recognition system used against him, including proprietary details like the technology’s underlying source code. The trial judge in Miles’s case applied that same Arteaga framework and granted the defense motion in full. The State responded by providing only limited discovery and then sought leave to appeal that ruling, an appeal the Appellate Division ultimately declined to hear.
The New Jersey Supreme Court took up the matter and issued a unanimous decision, authored by Justice Douglas Fasciale, that both reversed and affirmed different portions of the trial court’s original order. Central to the Court’s reasoning was a rejection of applying Arteaga mechanically to every case simply because facial recognition technology happened to be involved. Instead, the justices determined that what a defendant is entitled to receive through discovery depends heavily on the specific facts and circumstances of that particular case, rather than defaulting automatically to the same fixed thirteen item list regardless of context. As Justice Fasciale explained, this kind of discovery cannot be reduced to a rigid checklist, since what a given defendant is entitled to necessarily varies based on the specifics of their individual case.
Applying that case specific framework to the facts before it, the Court determined that the State was required to produce two particular categories of discovery. The first covers information identifying exactly which facial recognition tools and materials investigators actually used during the case, information the Court found directly relevant to building a fuller record regarding how reliable that particular technology actually is in this context. The second covers discovery related to precisely how the State used those specific tools and materials in the course of prosecuting Miles, information the justices found could prove relevant for challenging the reliability of witness identifications, scrutinizing the overall investigation, or even supporting an argument that someone other than Miles committed the crime. According to the Court’s opinion, this baseline information will, in the great majority of cases, represent the minimum a defendant needs in order to safeguard their constitutional right to a fair trial, even though the justices declined to establish a fixed checklist covering every possible scenario.
Where the Supreme Court parted ways with the trial judge’s original ruling involved the more sensitive, proprietary side of the facial recognition technology itself. The justices found that compelling the State to hand over proprietary information, including the actual source code underlying the software, was premature at this stage of the proceedings. Justice Fasciale explained that whether this kind of proprietary information ultimately needs to be disclosed is a question that has to wait until the case develops a more complete factual record, noting that even Miles’s own defense team had acknowledged they did not yet know whether that proprietary information would actually prove necessary to mount his defense.
Throughout its opinion, the Court was careful to acknowledge the genuinely unique challenges that come with seeking discovery of proprietary technology, recognizing that facial recognition software often involves commercially sensitive trade secrets that companies developing the technology have a legitimate interest in protecting. At the same time, the justices made clear that protecting those proprietary interests cannot come at the expense of a defendant’s constitutional right to challenge evidence used against them, striking a careful balance between those two competing concerns rather than resolving the tension entirely in favor of either side.
The Miles decision carries meaningful implications well beyond this single Jersey City case, since it establishes the framework New Jersey courts will now apply anytime facial recognition technology plays a role in a criminal investigation. Rather than granting defendants automatic access to a fixed list of thirteen categories of information simply because their case happens to involve facial recognition software, courts must now conduct a genuinely individualized analysis in each case, weighing what specific information is actually relevant to that defendant’s circumstances against what’s truly necessary to guarantee a fair trial. As facial recognition technology continues expanding its role in law enforcement investigations across New Jersey, this ruling gives both prosecutors and defense attorneys a considerably clearer, though still fact dependent, roadmap for how these discovery disputes are likely to be resolved going forward.
