scan_0031.jpg will not find you the photo you need at a deposition.

The photos arrive from every direction: a client's phone camera roll, a collision center's emailed ZIP, a responding officer's body-cam screenshots, an insurance adjuster's inspection set, and a stack of scanned prints from the ER visit. The files all end up in the same folder. They all have names like DSC_1142.jpg, IMG_4467.HEIC, scan_0031.jpg, and photo (3) copy.jpg. The metadata timestamps disagree with each other and sometimes with the incident date.

You know what is in the folder. The problem is that nobody else does, and neither does anyone in your office who wasn't the one who downloaded it. When opposing counsel takes the deposition and asks about the photo showing the left rear quarter panel, you need to find it in under ten seconds, not spend three minutes scrolling an unmapped grid of thumbnails while the room watches.

Why nondescript filenames are a real operational problem

The damage is not just inconvenience at a deposition. It runs through the whole file:

• NRCP 16.1 disclosure preparation takes longer. Identifying and describing each exhibit for the initial disclosure requires someone to open every image and write a description by hand. On a file with two hundred photos, that is a half-day of clerical work that has to happen before the disclosure deadline.
• Gaps are invisible. A named folder tells you at a glance whether you have photos from the ER visit, the scene, and the vehicle. A folder of camera codes tells you nothing until you have opened every file.
• Exhibit assembly is slower and more error-prone. When the time comes to pull together the trial exhibit book, someone has to match each image to the list description while looking at a screen full of DSC numbers. One transposition and the wrong photo goes behind the wrong label.
• Handoffs break. When the associate who downloaded the photos leaves the matter, the institutional knowledge about what each file shows leaves with her. A named folder does not require institutional knowledge.
• The client's narrative gets lost. A photo named "2024.05.01 Photo - bruising to left forearm.jpg" tells the story of the injury the moment you see it in the file. A photo named IMG_4467.jpg tells you nothing. Twelve months from trial, the narrative matters.

What good photo naming looks like

The convention that holds up in practice is simple: a date prefix in YYYY.MM.DD format so files sort chronologically, a category label that is the same across the file (Photo, Scan, Video Still), and a plain-language description of what the image actually shows.

The result is a folder you can scan:

• 2024.05.01 Photo - damaged front bumper.jpg
• 2024.05.01 Photo - skid marks on road surface.jpg
• 2024.05.02 Photo - bruising to left forearm.jpg
• 2024.05.02 Photo - laceration above right eyebrow, sutured.jpg
• 2024.05.15 Photo - vehicle interior, deployed airbags.jpg
• 2024.06.30 Photo - healing scar, right forearm.jpg

A folder like that is scannable in thirty seconds. The paralegal building the NRCP 16.1 exhibit list can write descriptions without opening each file. The attorney pulling the exhibit at trial can find the airbag photo without scrolling past a hundred unlabeled thumbnails.

NRCP 16.1 and the exhibit identification obligation

Nevada Rule of Civil Procedure 16.1 requires a party to disclose, without awaiting a discovery request, a copy or a description by category and location of all documents, electronically stored information, and other items that the disclosing party may use to support its claims or defenses. For a PI plaintiff, that means identifying the photographs. A description like "photographs of plaintiff's injuries and vehicle damage" is technically responsive but does nothing to help the other side or the court navigate two hundred images. Courts and opposing counsel both notice when a photo list is a placeholder.

Descriptive filenames make the exhibit list substantive without extra work: the name of the file is the description. That is not a legal strategy, it is basic organization, but it is organization that takes hours to produce by hand and takes minutes when a machine proposes the names first.

This is general information about how descriptive naming connects to disclosure obligations. Verify the current text of NRCP 16.1 and any local rules, including Eighth Judicial District Court Rules, before relying on it for a specific matter.

The privacy point: injury photos are sensitive

A personal injury file contains some of the most sensitive images a law firm handles: close-up photographs of injuries, medical records reduced to photos, surgical sites, bruising patterns. A family law file may contain similar images used as evidence of domestic violence or abuse.

Sending those images to an outside AI vision service, even a reputable one, requires you to think carefully about your confidentiality obligations under Nevada RPC 1.6 and the firm's data-handling representations to the client. Cloud AI vision APIs process images on remote servers. The firm's client did not agree to that when he or she retained you.

DilloLex's vision model runs on the appliance in your office. The images are read by the on-box model and named on the appliance. No client image is transmitted to any outside AI service as part of this process. The appliance still connects to the firm's cloud storage for file retrieval and delivery; what it does not do is route client images through an external AI processing service for analysis.

That distinction matters for your RPC 1.6 analysis and for how you describe your data practices to clients.

How DilloLex helps with photo naming

The workflow is straightforward:

1. Drop the photo folder into the matter. Images from the client's camera roll, the adjuster's ZIP, scanned prints, anything the firm has collected goes into the matter folder the same way it always would.
2. The on-box vision model reads each image. The Qwen3-VL model, running entirely on the appliance in your office, examines what each photo shows: vehicle damage, anatomical location of injuries, scene elements, document types, and so on.
3. It proposes a descriptive name. Each image gets a proposed filename in the standard format: date prefix where a date is determinable, category label, plain-language description of what the image shows.
4. Low-confidence items go to the review queue. Images where the model is uncertain, blurry photos, close-ups without enough context, images where the date cannot be determined reliably, are flagged and held in a _Needs Review folder for a staff member to name or correct.
5. Staff reviews and approves. The names are proposals. A paralegal or legal assistant reviews the batch, adjusts any name that is wrong or incomplete, and confirms the set before it is used in disclosures or exhibit assembly. No name is final until a human has said so.

The attorney never has to name a photo. The paralegal reviews a proposed list rather than building one from scratch. The result is a named folder that accurately reflects the matter and is ready for disclosure, exhibit assembly, and the deposition where someone will ask you to find the bumper photo in under ten seconds.

What the model gets right and where it has limits

An on-box vision model reads images well when the subject is clear and the resolution is adequate: vehicle damage, visible injuries, scene photographs, documents, and medical imaging where the anatomy is identifiable. It proposes sensible names for the large majority of a typical PI photo set.

It has limits. A blurry close-up of skin tissue without context may not be nameable with confidence. An image that contains identifying information, a face, a document with a name on it, gets the description it can give and goes through the same staff review. A photo taken in poor light may get a low-confidence flag. The review queue exists precisely because the model is not infallible; every uncertain item waits for a human before it is used. Your staff sets the standard, and the model's proposals are only as final as your review process makes them.

This is a tool that assists your staff. It does not replace the paralegal's judgment about what a photo shows or the attorney's decision about which photos go into the exhibit book.

The same named set is ready for Bates stamping and exhibit assembly

Once the photos have descriptive names, the rest of the pipeline is faster. When you are ready to produce the set, DilloLex applies the Bates sequence across the numbered photos in the order you specify, consistent with the same stamping process described in the Bates numbering guide. When you are assembling the trial exhibit book, the named photos slot into the exhibit list without a separate identification step. When you are preparing the PI demand package, the same folder of named images is already organized well enough to pull from directly.

Descriptive naming is not a standalone project. It is the first step in an organized file, and the organization carries forward through every subsequent step that touches the photos.

Common questions

Does NRCP 16.1 require me to name my photos descriptively?

NRCP 16.1 requires you to disclose documents and other items you may use to support your claims or defenses, either by producing them or by providing a description by category and location. It does not dictate a filename convention. Descriptive naming makes your disclosure list substantive and your file navigable; neither outcome is required by the rule's text, but both reflect the kind of organized practice courts and opposing counsel notice. Verify the current text of NRCP 16.1 and any applicable local rules before relying on this summary for a specific matter.

Can I send injury photos to a cloud AI service to get names for them?

You can, but you should analyze the question under Nevada RPC 1.6 before doing so. Transmitting client photographs to a third-party cloud service for processing means that service handles client content on remote infrastructure. Whether that satisfies your confidentiality obligations depends on the service's terms, your engagement letter with the client, and your firm's data-handling practices. The practical answer is that running the vision model on hardware the firm owns, in the firm's office, avoids the analysis entirely for that step.

What if the photo has no reliable date?

If the model cannot determine a reliable date from the image itself or from the file's metadata, it flags the item as low-confidence and holds it in the review queue. A staff member can then supply the date from context, such as the ER visit date, the incident date, or the adjuster's inspection date, and confirm the name. No date is invented; items without a reliable date either get confirmed by staff or are left undated in the filename until the date is known.

How do I handle photos that show identifying information, like a face or a name on a document?

The vision model reads and names the image based on what it shows. If the image contains a face or other identifying information that may need to be redacted before production, the redaction decision belongs to the supervising attorney. DilloLex's redaction tool can scan the named photo set for PII and flag items for attorney-authorized redaction before any photo goes out in a production or demand package. The naming step and the redaction step are separate; naming does not affect what is or is not redacted.

Does this work for family law files that include photos as evidence?

Yes. The same on-box vision naming process applies to photographs in a family law matter, including images used as evidence in domestic violence, abuse, or custody proceedings. The privacy point is the same or stronger: images in a family law file are at least as sensitive as a PI injury photo set, and the same analysis under Nevada RPC 1.6 applies to how those images are handled. No client image from a family law matter is transmitted to an outside AI service as part of the naming process.

A note on accuracy

This guide is general information for Nevada legal professionals. It is not legal advice, it does not create an attorney-client relationship, and it is not a substitute for the current text of any rule. NRCP 16.1, the Nevada Rules of Professional Conduct, and the Eighth Judicial District Court Rules are amended from time to time; verify the current text of any rule referenced here before relying on it for a specific matter or client.

See DilloLex name a photo set in your office