What the Brady Rule Actually Requires (And Why Violations Keep Happening)

In 1963, the Supreme Court held in Brady v. Maryland that prosecutors must disclose favorable, material evidence to the defense. That duty is a constitutional floor, not a discovery courtesy: it applies whether or not the defense asked for it, and it covers both exculpatory evidence and impeachment material. Giglio v. United States (1972) extended the rule to undisclosed deals for testimony, and Kyles v. Whitley made clear that the prosecution answers for favorable evidence held by anyone on the government's side, including police, even if no prosecutor ever saw the file.

Today, the operative test comes from Strickler v. Greene. The defense must show that the evidence was favorable, that the State suppressed it, and that the suppression prejudiced the outcome. All three must be true. Notably, intent is not an element: the violation stands whether the prosecutor acted in good faith or bad.

So why do violations keep happening? The largest empirical study of the problem, analyzing 386 adjudicated Brady violations from 2004 to 2022, offers a sobering picture. Prosecutors acted in bad faith in 66% of cases, rising to 74% in homicide cases, where nearly half of all violations occurred. But the same study found that in 51% of successful claims, the Brady material never even made its way to the line prosecutor. The system is failing at the handoff, not just at the decision to withhold.

That handoff problem has gotten worse, not better, as discovery has gone digital. Aaron A. Pelley of Gleam Law put it bluntly: if the defense cannot reasonably find or understand the favorable evidence, the State may have complied on paper while undermining Brady in practice. Favorable evidence now hides in plain sight inside large file dumps, unsearchable formats, metadata, and missing files. A jail call buried in a folder of 40 hours of bodycam footage is disclosed in name only.

The structural drivers compound the problem. Cook County prosecutors carry 300 or more open cases at a time, and 15% to 33% of prosecutor positions sit vacant in large district attorney offices. Only one state requires prosecutors to maintain Brady lists at all, and disclosure practices often depend on the changing whims of elected officials.

For the defense, the practical takeaway is to stop relying on the prosecution's process. Make Brady-specific requests early, narrow, and on the record, because generic requests get generic productions. Treat Brady list research as its own workstream, separate from general discovery review. And build a review workflow that actually surfaces inconsistencies across the full dump: bodycam against police reports, jail calls against witness statements. That is the work Evidence Copilot is built to support, with timestamped transcripts and citation-backed review that flag report-versus-footage conflicts on ingest, before they become appeal issues instead of trial exhibits.

The Defense Workflow That Surfaces What Prosecutors Miss

If the State's disclosure system is "a jigsaw puzzle with a thousand tiny pieces" where no one is in charge, waiting for a complete Brady file is a losing strategy. The defense workflow has to assume material will slip through, then build the record that catches it when it does.

That assumption is grounded in the data. In the largest empirical study of adjudicated Brady violations (386 cases, 2004–2022), the material "never even made its way to the line prosecutor" in 51% of successful Brady claims. A generic "produce all favorable evidence" letter feeds right into that failure mode. Generic requests get generic productions; specific ones, like informant deals, prior inconsistent statements, lab results, and officer disciplinary records, are harder to bury and easier to prove were ignored.

So the working workflow looks like this:

• Request early, narrow, and on the record. A dated, itemized letter creates a record the State cannot later claim it never received. Brady applies whether or not you asked, but the request is what makes suppression provable.
• Run Brady-list research as its own workstream. Only one state requires prosecutors to maintain Brady lists, and few jurisdictions regulate them at all, so you cannot rely on the prosecution to surface officer misconduct. Request disciplinary and IA files independently and document every ask.
• Track every inconsistency across the record. Brady claims often live in gaps between what a witness said at different times. Compare bodycam against police reports, jail calls against witness statements, and 911 audio against officer narratives, and log who said what and when.
• Document chain of custody from day one. Show when evidence existed, who held it, and when it should have been disclosed. Gaps work against whoever cannot account for them.

The inconsistency tracking matters more than it used to, because the digital evidence explosion has made favorable evidence "hidden in plain sight" in large file dumps, unsearchable formats, and missing files. As defense-side commentary puts it: if the defense cannot reasonably find or understand the favorable evidence, the State may have complied on paper while still undermining Brady in practice. With 45% of adjudicated violations involving suppressed impeachment material for a critical prosecution witness, the cross-report sweep is where most of the value sits.

This is where searchable, timestamped review earns its keep. Evidence Copilot was built for this step of the workflow: upload the discovery dump and get diarized transcripts, OCR'd document text, and candidate inconsistencies flagged between report and bodycam or statement and call audio, each answer citing the exact second of video or page of PDF. Every AI output is marked pending human review, so the attorney, not the tool, makes the Brady call. A structured Brady list research workflow turns the same discipline into a repeatable process for officer-records requests.

One more reason to start now rather than at the pretrial deadline: prosecutors acted in bad faith in 66% of adjudicated violations, rising to 74% in homicide cases. A contemporaneous record of requests and productions is what preserves that issue for trial and for post-conviction, long after memories fade.

Building a Searchable, Citation-Backed Record Before Trial

The most dangerous Brady violation isn't the file the prosecutor hides. It's the exculpatory statement sitting in hour 31 of bodycam footage that nobody had time to watch. As practitioners have noted, the digital evidence explosion lets favorable evidence hide "in plain sight" inside large dumps, unsearchable formats, and metadata. If you can't find it, you can't litigate it.

The numbers explain why you can't outsource this to the State. In the largest empirical study of adjudicated Brady violations (386 cases, 2004–2022), 51% of successful claims involved material that never even reached the line prosecutor. And 45% of violations involved suppressed evidence that could have impeached a critical prosecution witness. That impeachment material is exactly what a searchable record surfaces first.

A grounded evidence workspace changes the math. Upload the discovery dump and everything becomes findable: bodycam, interrogation footage, jail calls, 911 audio, and scanned police reports all get timestamped, diarized transcripts or OCR'd text. Every AI answer cites the exact video second or PDF page it came from, so "the officer said otherwise" becomes a clickable moment you can pull up in a hearing. On ingest, the system compares each new file against earlier evidence and flags candidate inconsistencies, report-versus-bodycam, statement-versus-call audio, before anyone asks. Every flag is marked AI-generated and pending your review. The originals stay immutable.

That record feeds three workstreams directly:

• Suppression prep. A Miranda Timeline and suppression sweep drafts a motion outline with both sides quoted and cited, so your consent, search-scope, or Miranda argument is built on exact moments, not memory.
• Cross-exam drafting. From a flagged inconsistency, you get suggested questions tied to quoted sources, with no credibility conclusions. You make the judgment call; the citation backs it.
• Exhibit building. The Exhibit List Builder ties each exhibit to the real file, and the export produces a citation-indexed exhibit pack, so what you argue in court matches what's in the record.

This is why Brady list research workflow belongs in your case setup, not your appeal posture. When inconsistencies surface before trial, they become cross-examination and motions. When they surface after conviction, they become years of post-conviction litigation. One attorney quoted in the research put it plainly: had today's technology existed at the time, inconsistent police statements "would have been flagged immediately as part of discovery." Your record should be the one that's searchable, timestamped, and citable, because the prosecution's won't be built for you.

What to Do When You Find a Brady Violation

Finding a Brady violation is only half the job. The other half is preserving a record clean enough that a judge can actually do something about it, because remedies evaporate when your documentation is thin.

Start with the pre-trial record. If you catch suppression before trial, ask the court for a disclosure order, a continuance to investigate the new material, or exclusion of the tainted testimony. Judges can order disclosure, delay proceedings, or exclude testimony, and the choice among those remedies depends heavily on how well you can show when the evidence existed, who had it, and when it should have been disclosed (practitioner guidance). Gaps in the timeline work against whoever can't account for them, so log every request and production as it happens, not from memory at the hearing.

That contemporaneous record matters more than most attorneys assume. In the largest empirical study of adjudicated Brady claims, 386 cases from 2004 to 2022, prosecutors acted in bad faith in 66% of cases, rising to 74% in homicide prosecutions where a prosecutor was the suppressor (the McAward study). Bad faith findings follow a paper trail: specific written requests the State ignored, disclosure deadlines it blew, representations it made to the court that turned out wrong. Generic requests get generic productions, and they also give the prosecution an easy out later. Narrow requests for informant deals, prior statements, lab results, and officer disciplinary records are harder to bury and easier to prove were ignored.

If the violation surfaces post-conviction, the remedies shift to a new trial, a vacated conviction, or dismissal of charges. Here the record you built at trial becomes the appellate record, and 51% of successful Brady claims involved material that never even reached the line prosecutor, which makes chain-of-custody documentation of the discovery dump your strongest exhibit.

Manage expectations on civil suits. Claims against prosecutors rarely succeed because of absolute immunity for conduct tied to their prosecutorial role. Bar discipline, contempt findings, and public reprimands happen, but they're rare relative to how often violations surface. The practical remedy is almost always the criminal case itself, which is another reason to fight for disclosure orders early rather than plan on accountability later.

The workflow that supports all of this is unglamorous: timestamped transcripts, OCR'd documents, and a running log of what was requested and produced. Tools like Evidence Copilot are built for exactly that, surfacing candidate inconsistencies between a police report and bodycam footage with citations to the exact moment, so your Brady record is assembled as you review rather than reconstructed under deadline. For officer-mismaterial specifically, a structured Brady list research workflow keeps the Giglio workstream separate from general discovery review, since only one state even requires prosecutors to maintain Brady lists and you can't count on the other side to volunteer misconduct records.