Body Camera Footage Redaction: FOIA Fines & Rulings [2026]

For a decade, the standard answer from an under-resourced records unit was some version of "we'd release it, but we can't redact it." In 2026 that answer started costing money. Connecticut's Freedom of Information Commission fined the City of New Haven $1,250 over a body-camera request that had sat unreviewed in a queue. A DC court ordered the Metropolitan Police Department to hand over more than a thousand hours of body-worn camera video after rejecting its plan to blur nearly everyone in frame. Colorado's Court of Appeals told Boulder it could not bill a magazine thousands of dollars for the labor of preparing bodycam footage. Illinois' Attorney General voided a per-minute redaction fee outright.

The pattern is consistent: the third-party privacy interest in body camera footage is real, and it is exactly why redaction exemptions exist — but the legal remedy is to redact and release, not to withhold or to bill the requester for the difficulty. This guide covers what actually has to be blurred in a body camera release, why frame-by-frame manual redaction is the bottleneck that creates these backlogs, and how motion-tracked automated blur changes the cost equation for an agency without an enterprise redaction budget.

Can a Police Department Deny a Bodycam Request Because It Can't Redact the Video?

No — and the authority saying so is getting harder to ignore. Public records statutes are built around segregability: the agency withholds the exempt portion and releases the rest. An inability to perform the redaction is an operational constraint on the agency, not a statutory exemption, and no state open-records act contains a "we lack the software" carve-out.

Illinois offers the cleanest recent statement. In Public Access Opinion 25-014, issued December 3, 2025, the Attorney General's Public Access Bureau addressed the Mattoon Police Department, which had told a requester that 4.5 hours of body-worn and dashboard camera footage "must have redactions made" before disclosure and demanded $696.60 in prepayment at a rate of "$2.58 per minute to redact." The binding opinion held the fee improper and directed the department to produce the footage "subject to appropriate redactions, and assessing him a fee of no more than the actual cost of purchasing a recording medium."

Read that holding carefully, because it cuts both ways. The department was not excused from redacting. It was told to redact, release, and eat the cost.

Illinois agencies have reportedly tried the blunter version of the argument too — the City of Harvey is frequently cited as having denied a body camera request on the grounds that it lacked the means to blur faces and protect the identities of third parties, with the resulting FOIA suit settling. We could not locate a primary court record or contemporaneous news report for that settlement, so treat the specifics as unconfirmed; the durable authority is the binding opinion above, and it points the same direction.

What Did the Connecticut FOI Commission Actually Fine New Haven For?

It fined the city $1,250 for what the Commission called "a clear disregard" of the public's rights under the FOI Act, after a body camera request went nearly three years without production. The final decision in Docket #FIC 2025-0138, dated February 25, 2026, names then-Chief Karl Jacobson, the New Haven Police Department and the City of New Haven as respondents.

The procedural history is the part every records officer should read. The complainant's original request covered 911 audio, CAD reports and body-camera footage for three case numbers. An earlier Commission decision in November 2024 had already found the city violated the FOI Act's promptness provisions — and in that first proceeding, the Commission noted, the department "had yet to review the body-camera footage as it remained 20th in their queue of requests for body-camera footage."

That single line is the whole story of the redaction backlog in legal form. Because nobody had watched the video, the department could not identify what was exempt, could not assert an exemption, and could not carry its burden of proof. The Commission found the respondents "failed to assert any claim of exemption or provide any evidence as to the applicability of any exemption," ignored a subsequent order to submit a supporting affidavit, and then failed to brief the issue again in the penalty proceeding. The New Haven Independent reported the fine and the city's plan to appeal, and CT Examiner covered the appeal itself in July 2026.

A backlog does not just delay a release. It destroys the agency's ability to defend the exemptions that protect the bystanders in the footage.

Why Did DC Police Lose the January 6 Bodycam Case if Bystander Privacy Matters?

Because it proposed to blur almost everyone, and a court read that as withholding by another name. Judicial Watch's account of its DC FOIA suit (No. 2024-CAB-003453, filed June 2024 over an August 2021 request) describes an April 2026 ruling in which the court rejected the Metropolitan Police Department's position that it could withhold the faces and voices of everyone except law enforcement, and rejected the department's estimate that production would cost more than $1.5 million. The court found any privacy interests minimal against the overwhelming public interest in disclosure. In July 2026 the department produced 1,627 videos totalling over 1,000 hours of body-worn camera footage.

There is a real lesson here that has nothing to do with the politics of January 6. Blanket redaction is legally fragile. When an agency proposes to obscure every face and every voice in a release, it is not performing segregability analysis — it is asserting a categorical exemption that the statute does not grant, and it is inviting a court to conclude that the privacy interest was never seriously weighed. Targeted redaction of the people who are actually exempt is both cheaper to defend and cheaper to produce.

The cost argument fared no better. A $1.5 million production estimate is, in practical terms, an admission that the department's redaction workflow scales linearly with headcount. Courts have grown unsympathetic to that framing precisely because automated tooling now exists.

Can an Agency Charge Me for the Cost of Redacting Body Camera Video?

It depends on the state, and 2026 has been an unusually active year on this exact question. Four jurisdictions illustrate the whole spread:

Colorado's case is the sharpest. Boulder demanded that Yellow Scene Magazine pay more than $8,000 for body camera footage of a December 2023 shooting in which officers killed Jeanette Alatorre. The magazine sued, won at the district court, and on April 9, 2026 the Colorado Court of Appeals affirmed that agencies cannot charge fees for bodycam footage of alleged officer misconduct under the state's Law Enforcement Integrity Act, noting the "conspicuous absence of a fee provision" in the statute. Boulder, joined by the Colorado Municipal League and the County Sheriffs of Colorado, petitioned the Colorado Supreme Court in June 2026, arguing the ruling imposes "enormous ongoing costs" on agencies.

Ohio went the other way. Governor Mike DeWine signed House Bill 315 in January 2025, permitting departments to charge up to $75 per hour of video requested, capped at $750, explicitly to recoup the labor of reviewing and obscuring footage.

The Nevada Case: Fees as a De Facto Denial

The freshest thread runs through Las Vegas. On July 30, 2026, the ACLU of Nevada, a Nevada law firm and two news organizations including The Marshall Project filed suit in Clark County District Court against the Las Vegas Metropolitan Police Department over public-records fees. Per The Nevada Independent's reporting, the ACLU received a bill of nearly $7,000 for body camera footage, backed by an estimate of over 80 hours of officer work to assemble the records. The complaint argues that charging for officer time violates Nevada law because those employees are salaried regardless of assignment.

Whatever the outcome, the mechanism is the point: when redaction labor is the dominant cost of a records release, the fee schedule becomes the access policy. Reduce the labor and the fee fight largely disappears.

Why Is Manual Redaction Such a Severe Bottleneck?

Because targeted blurring costs roughly ten minutes of skilled labor for every minute of footage, per subject. That figure is not a vendor estimate — it comes from the Seattle Police Department's own stopwatch study.

Seattle's revised Body Worn Video Redaction Cost Study, published October 2024, breaks redaction into three types and measures each:

Run the arithmetic on a realistic scene. A 20-minute traffic-stop clip with two bystanders on the sidewalk, one passenger and one non-subject vehicle plate has four tracked objects. At 10 minutes of work per minute of footage per object, that is 800 minutes — over 13 hours — for a single 20-minute clip. Multiply by the four to six officers who responded and you have a week of one specialist's time for one incident.

Now recall New Haven's queue. Twentieth in line, at that throughput, is measured in years. The backlog is not negligence; it is arithmetic. Which is precisely why "we can't keep up" stopped persuading commissions and courts once faster tooling became commercially ordinary — the enterprise redaction tooling comparison walks through what the market actually offers at each budget level.

What Actually Has to Be Blurred in a Bodycam Release?

The exempt categories are narrower than a blanket blur but broader than most people expect. Federal FOIA exemptions b(6) and b(7)(C) under 5 U.S.C. § 552 protect personal privacy generally and personal privacy in law enforcement records specifically; nearly every state statute carries an analogue.

In practice, a release checklist covers:

• Faces of uninvolved bystanders and witnesses who happened to be in frame
• Minors, in essentially all circumstances
• Victims of sensitive offences — domestic violence, sexual assault — absent consent
• Confidential informants and undercover personnel
• License plates of non-subject vehicles, which are trivially traceable once published
• Identity documents and paperwork held up to the camera: driver's licenses, insurance cards, prescription labels, medical forms
• In-car computer and phone screens showing criminal history, dispatch data or NCIC returns
• Interiors of private homes and medical facilities, where Seattle's study uses full-screen blur rather than targeted blur
• Audio naming exempt individuals, reading out dates of birth, or capturing witness statements

Two of those categories are routinely missed. Text redaction is the first: a license held up to a body camera for two seconds will be legible on a 4K frame, and OCR-driven automatic sensitive text redaction catches document and screen content that a face detector ignores. Audio is the second — visual redaction alone leaves a soundtrack full of names, and voice anonymization handles the portions that must be obscured rather than cut.

Is a Blur Actually Strong Enough, or Can It Be Reversed?

Weak blur is reversible, and this matters legally as well as technically. Light Gaussian blur and coarse pixelation both preserve enough structure for reconstruction techniques to recover recognizable features, which is why regulators increasingly distinguish obscured from anonymized. Our breakdown of why weak blur fails forensic de-pixelation tests covers the threshold, and the guide to what "redacted" actually means explains why a redaction that can be undone is not a redaction at all.

The other failure mode is temporal. A blur applied to frame 1 and not frame 340 is a full identity disclosure, and handheld body camera footage is the worst case for this — the wearer turns, the subject walks, the camera shakes. Temporal consistency in video anonymization is the difference between a defensible release and an embarrassing one.

How Does Motion-Tracked Automated Blur Change the Math?

It collapses the per-object cost from "ten minutes per minute of footage" to a single detection pass plus human review. That is the entire difference between a queue that grows and a queue that shrinks.

✅ Motion-Tracked Face Blur

BGBlur detects faces and tracks them across frames, so a bystander who walks behind a parked car and reappears stays covered without a manual keyframe. This is the feature that matters most for body camera footage specifically: the camera is strapped to a moving human, so nothing in frame is static and every subject needs continuous tracking rather than a fixed mask.

✅ Automatic License Plate Blur

Plates on non-subject vehicles are one of the most commonly missed items in a release, partly because they enter and leave frame constantly during a walk-up. Automated license plate detection and blur handles the plates in a scene in the same pass as the faces.

✅ Selective Object and Text Redaction

Not everything exempt is a face or a plate. Text-prompt selection lets you designate a specific document, screen, tattoo, house number or item that needs obscuring, which is what a segregability analysis actually requires — you redact the exempt thing, not the whole frame.

✅ Batch Processing for Multi-Clip Releases

A single incident generates footage from every responding officer. Batch processing queues the whole set rather than forcing one-file-at-a-time handling, which is where multi-officer releases usually stall.

✅ Browser-Based Processing with 24-Hour Deletion

Processing runs client-side in the browser, and files are deleted within 24 hours. For records officers uneasy about pushing evidentiary video into a third-party cloud, and for journalists who received footage under an embargo, that architecture is a meaningful difference from an upload-and-hope pipeline.

How to Redact a Bodycam Clip for Release: A Practical Workflow

Step 1: Watch the footage before you promise a timeline. New Haven's penalty traces directly to footage nobody had reviewed. You cannot assert an exemption you have not identified, and a five-minute skim tells you whether a clip has two tracked objects or twelve.

Step 2: Build a written redaction log per clip. List each exempt item, the statutory basis, and the timecode range. This log is what you attach when a commission or court asks you to justify the redaction — and it is what turns a blanket blur into a defensible segregability analysis.

Step 3: Run automated detection for faces and plates. Upload the clip, enable face blur and license plate blur, and let the detection pass run. Verify strength: use a heavy blur or solid mask rather than a light smudge, for the de-pixelation reasons above.

Step 4: Add selective redactions for documents, screens and specific individuals. Use text-prompt or region selection for the items automated detection is not designed to catch — an MDT screen, an ID held to camera, a house number. Full-screen blur is appropriate for private-residence and medical-facility interiors.

Step 5: Review the output frame by frame at transitions. Check the moments where subjects enter and exit frame, where the camera whips, and where lighting changes. These are where tracking is most likely to drop, and they are the only part of the workflow that still genuinely needs a human.

Step 6: Release with the log, retain the original. Produce the redacted copy, keep the unredacted master under your retention schedule, and disclose what was redacted and why. Agencies that document the redaction rationale almost never end up in a penalty proceeding.

For a walkthrough focused purely on the mechanics, see our step-by-step guide to blurring body cam video footage.

Who Actually Benefits — and Where BGBlur's Scope Ends

Being honest about fit matters more than a feature list here, because this is a domain with real certification requirements.

Small and mid-size agencies without an enterprise redaction budget. A department with one records clerk and no six-figure contract is the profile most exposed to fines and adverse rulings. BGBlur's Business tier is $29/month with batch processing and 4K export — a rounding error next to a $1,250 civil penalty, let alone a fee-shifting order.

Journalists and newsrooms who receive footage. Once footage is in your hands, the exemptions no longer bind you but the ethics do. Blurring bystanders, minors and plates before publication is the single most effective way to avoid a takedown demand.

Requesters, attorneys and advocacy organizations preparing exhibits or public releases from footage they obtained, where the third parties in frame never consented to anything.

Where the scope ends: BGBlur is a redaction tool, not a CJIS-certified evidence management system. It does not provide chain-of-custody logging, evidence retention, CJIS-compliant storage or an audit trail for Criminal Justice Information. Agencies whose policy or state law requires CJIS certification of every system touching footage should use a certified platform for CJI-bearing video — the enterprise video redaction comparison covers those options. BGBlur fits non-CJI footage, backlog clearance, and everyone downstream of the agency.

What to Do If Your Request Was Denied Because "We Can't Redact It"

1. Ask for the denial in writing, with the statutory exemption cited. "We lack the capability" is not an exemption, and an agency that has to write it down often reconsiders.
2. Cite the segregability obligation. Most state acts require release of non-exempt portions of a record; point to the specific provision.
3. Check whether your state permits redaction fees at all. In Illinois it does not. In Colorado it does not for misconduct footage. A fee demand may itself be unlawful.
4. Narrow the request strategically, not passively. Asking for one officer's camera for a ten-minute window cuts the redaction burden by an order of magnitude and removes the agency's best excuse.
5. Appeal to the state commission or public access counselor. The New Haven penalty and the Illinois opinion both exist because someone escalated rather than giving up.
6. Offer the tooling answer. Agencies sometimes genuinely do not know that motion-tracked automated redaction costs $29 a month rather than $50,000 a year.

For the broader context of filming and publishing footage that captures people who did not consent, our guide to unauthorized filming and privacy protection and the coverage of body camera and drone footage at protests go deeper.

The Redaction Excuse Has Expired

The through-line across Connecticut, Colorado, Illinois, DC and now Nevada is not that courts have stopped caring about the bystanders in body camera footage. They care a great deal — that concern is the entire reason the privacy exemptions exist. What has changed is that decision-makers no longer accept redaction difficulty as a reason to withhold, delay indefinitely, or shift the cost onto the person asking. A $1,250 penalty, a fee ban, a voided per-minute charge and a thousand-hour production order all say the same thing in different words: redact and release.

For agencies, the practical takeaway is that the queue itself is now the liability. Footage nobody has watched is footage whose exemptions cannot be defended, as New Haven learned when its inability to review a clip stranded 20th in line became a finding of clear disregard. For journalists and requesters, the takeaway is that receiving footage transfers the privacy obligation to you in practice even when it does not in law.

Motion-tracked face blur, automatic plate detection, selective object and text redaction and batch processing turn a ten-minutes-per-minute manual task into a detection pass plus a review pass. That is not a marginal efficiency gain — it is the difference between a backlog that compounds and one that clears. Try BGBlur on a clip and see how long the review actually takes.