Article

Section 01

What happens to a photograph after it leaves your hands

A photograph rarely stays where it was first published. A portfolio image, a picture delivered to a client, a file licensed through a stock agency or a frame posted on social media can be copied within hours, then copied again from the copy. Each new appearance is a separate act of reproduction and public display, and each may be lawful or not depending on the licence behind it.

Most photographers learn about a reuse by chance: a client mentions it, a colleague sends a screenshot, a search turns up something unexpected. By then the use may have been online for months, the page may have changed, and the person who published it may no longer remember where the file came from.

Monitoring replaces chance with a routine. It does not create rights and it does not prove infringement on its own. What it does is give you an accurate, dated picture of where your work appears, so that every later decision rests on facts rather than recollection.

This article explains how images travel across the web, why timing matters, what search and monitoring tools can and cannot do, how to avoid treating licensed uses as problems, and how a finding becomes a documented review at Atlantic Image Defense LLC.

Section 02

How images spread: the usual channels

Unauthorised reuse follows recognisable patterns. Knowing them helps you decide where to look and how to read what you find.

• News and content aggregators republish articles from other sites, often with the original images, sometimes stripped of captions and credits.
• Blogs and small editorial sites take images to illustrate posts, frequently straight from a search engine results page, on the assumption that anything visible is free to use.
• E-commerce listings and marketplaces use photographs to sell products, accommodation or travel services, including pictures of places, interiors or food taken from a photographer’s site or a supplier’s catalogue.
• Social media accounts repost images, crop them or add text overlays, and the reposts are then picked up by websites that embed or download them.
• Scraper sites copy content automatically from other sources to generate traffic and advertising revenue, often across many near-identical domains.
• AI-generated or automatically assembled pages combine scraped text with images pulled from elsewhere; they can appear and disappear quickly and are often spread over rotating domains.

Two features recur across these channels. Copies multiply from copies, so the site you find is often not the first one to have used the image. And the further an image travels, the more likely its metadata and credit have been lost along the way, so the person who publishes it is less likely to know who the author is.

Neither feature changes the legal position. Under US law and in most other jurisdictions, each publication is assessed on its own, and the fact that an image was found elsewhere does not grant a right to use it.

Section 03

Why early detection matters

Web pages are not stable. A page can be edited, an image replaced, a site redesigned, a domain allowed to lapse. When a use is found late, the most useful evidence is often gone: the original layout, the context in which the image was shown, the surrounding text that made it commercial or editorial. Early detection lets the use be captured while it is still in its original state.

Timing matters for the chronology as well. A well-built file shows when the use was first observed, how long it remained online, whether it changed after contact, and whether it stopped. These points shape any later discussion. A use that has run for a long time on a commercial page is not assessed in the same way as one that appeared briefly on a personal blog, and only dated records can show which situation applies.

Under US law, dates also interact with registration. Statutory damages and attorney’s fees under 17 U.S.C. § 412 generally require timely registration, and whether registration was timely is a question answered with dates. Registration within five years of publication is prima facie evidence of validity under § 410(c). Monitoring does not replace registration, but a dated record of when a use was first observed helps counsel assess what is available in a given matter.

References17 U.S.C. §§ 412, 410(c)

For works first published outside the United States, registration is not needed before suing in the US under § 411(a), and the Berne Convention, art. 5(2), provides that the enjoyment and exercise of rights are not subject to any formality. The dates still matter, because they are the facts of the case.

References17 U.S.C. § 411(a) · Berne Convention

Early detection also helps everyone involved in practical terms. A website operator contacted soon after publishing an image can often check with its supplier, its designer or its own licence records while those are still easy to find. A late contact makes that harder for both sides.

Section 04

Reverse image search: what it does and where it stops

Reverse image search lets you submit a photograph and receive pages where the same or a visually similar image appears. It is the most accessible starting point, and it finds many obvious reuses, especially on public pages that search engines have indexed.

Its limits are worth understanding before relying on it:

• It only finds what has been indexed. Pages behind logins, new pages not yet crawled and sites that block search engines may not appear.
• Results change over time. The same search run next month can return different pages, and a page may drop out of the results without having been removed.
• Heavy crops, mirrored images, colour changes, text overlays and composites can reduce matching, particularly when only part of the frame is used.
• Visually similar results are not matches. A search may return photographs of the same landmark or product taken by someone else.
• A result is a pointer, not evidence. The thumbnail and link show that a match was detected; they do not record what the page looked like, when, or in what context.

For these reasons a reverse image search result should always be opened, checked and, if relevant, captured properly. A screenshot of a results page can help show how a use was found, but it is not a capture of the use itself.

Section 05

Continuous monitoring and its limits

Continuous monitoring repeats searches across the open web at intervals and flags new appearances of the images submitted to it. Its value lies in regularity: it notices new uses without depending on someone remembering to look, and it builds a history of when each appearance was detected.

Monitoring has limits of its own, and a credible file acknowledges them rather than glossing over them:

• Coverage is never complete. No system sees every page on the web, and closed platforms, apps and private groups are largely out of reach.
• The detection date is the date on which the system observed the use, not necessarily the date on which the use began. The page may have been online earlier.
• Automated matching produces candidates. Some will be the same image, some a different photograph of the same subject, and some a licensed use.
• Social media is particularly difficult: much content sits behind logins, platform terms restrict automated collection, and each platform runs its own reporting procedure.
• A page that was online when detected may be gone when someone returns to it. This is why capture at the time of detection matters.

It also helps to be selective about what is monitored. Images that are widely licensed, frequently shared or commercially attractive (travel, food, architecture, interiors, product and lifestyle work) tend to be reused more often than others, and they are the ones where a regular watch pays off. Images under an exclusive licence deserve attention too, because an unexpected appearance elsewhere may matter to your client as well as to you.

Monitoring is therefore a detection tool. What turns a detection into something usable is the capture and the review that follow it.

Section 06

False positives and licensed uses

Not every match is a problem, and treating every match as one quickly damages credibility. Two kinds of result need to be set aside before anything else happens.

False positives are results that are not your image at all: another photographer’s picture of the same building, a similar composition, a stock image of a comparable scene. They are excluded by direct comparison with your original file, looking at details that two separate photographs would not share, such as the exact position of clouds, passers-by, reflections or light.

Licensed uses are your image, used lawfully. They are common and should be expected, particularly by stock contributors and by photographers who license through agencies or directly to clients. A client may have sublicensed within its rights, an agency may have sold a licence that has not yet appeared in your statements, or a publisher may hold an editorial licence for a specific article. Signs worth checking include:

• a credit to a stock agency or image bank next to the image or in the page footer;
• a match with a sale recorded in your contributor or agency statements;
• an editorial context consistent with a licence you or your agency granted;
• a past or present client relationship with the business that published the image;
• a use on a page belonging to a client’s own customer, distributor or partner.

When one of these signs is present, the right step is to ask for the licence record, not to send a claim. A party that holds a licence can usually produce it, and the matter can then be closed. A credit to the photographer, on the other hand, is not in itself a sign of a licence: crediting the author is good practice, but it does not grant permission.

Some results remain uncertain after these checks: the image may be yours, but the page offers no clue about how it was obtained. Those results should be recorded as open questions, not as findings of infringement. They can be captured, kept in the log and reviewed again later, once more information is available from your agency, your clients or a further search.

Section 07

From a finding to a documented review at Atlantic

At Atlantic, a detection is not yet a case. It becomes one only after it has been captured and reviewed.

Each finding is recorded as an occurrence. The capture includes the full URL of the page, the URL of the image file itself, the date and time of capture, a full-page screenshot showing the image in context, and a copy of the image as served by the site. A SHA-256 fingerprint is computed for each evidence file, so that it can later be shown that nothing has been altered since capture. Every capture and every later action is entered in a chain-of-custody record.

The review then works through a series of questions:

• Is the image the rights holder’s photograph, as opposed to a similar one?
• Has the rights holder’s authority to act been verified for this image, through ownership, a licence chain or a mandate?
• Are there signs of a licence, such as an agency credit, a recorded sale or an editorial context?
• Who operates the site, and is the use commercial, editorial or personal?
• Is this occurrence linked to others on the same domain or on a network of related sites?

Uses that are clearly covered by a licence are closed without any claim. Where the review supports it, a Case File is opened with a reference of the form AID-YYYY-NNNNN; where a licence is plausible, Atlantic first asks for the licence record; a detection that the review does not support never becomes a Case File. The operator then receives access to a secure case portal where the evidence summary is visible and where they can reply, upload a licence, an invoice or other proof, or settle. The correspondence that follows, and the checks made before any notice to a host, are covered in separate articles.

Section 08

Repeated reuse by the same domain

Monitoring sometimes shows one domain using several of a photographer’s images, or using the same image again after an earlier use was taken down. This pattern needs careful handling, because it can mean very different things.

It can reflect a single source: a website template, a supplier’s image pack, or a designer who drew on the same folder for several pages. It can reflect a licence covering several images that the operator genuinely holds. Or it can reflect a habit of taking images without checking where they come from. The review does not assume the worst. It groups the occurrences, checks each one for signs of a licence, and records them together, so that the operator receives one coherent picture rather than a series of separate messages.

Where an image reappears after the operator removed it, the dates become particularly important. Removal ends the ongoing use but does not erase the earlier period of use, and a later republication is a new act, documented in its own right. Under US law, where infringement is found to be willful, a court may increase statutory damages under 17 U.S.C. § 504(c)(2). Whether any conduct is willful is for a court to decide, not for a monitoring record and not for Atlantic. What the record can do is show, accurately and without commentary, what was online and when.

References17 U.S.C. § 504(c)(2)

Repeated reuse across a network of related domains, which is common with scraper and automatically generated sites, is recorded in the same way: each domain, each URL, each date. Where the sites appear to share an operator or a host, the file states it only once it has been verified.

Section 09

Keeping your own monitoring log

Whether or not you work with Atlantic, a simple monitoring log of your own is worth keeping. It gives you a memory that does not depend on search results staying the same, and it makes any later file quicker and more accurate to assemble.

A useful log records, for each finding:

• the date you found it and how (which search, which tool, who told you);
• the page URL and, where you can obtain it, the image URL;
• which of your images it is, with your file name or catalogue reference;
• whether a capture was made, and where it is stored;
• what you checked about licensing (agency statements, client records) and what you found;
• what you decided: no action, a licence request, or submission for review;
• any later change you observe: image removed, page edited, image reused elsewhere.

Keep the log factual. Write what you saw, not what you believe the other side intended. Comments such as “obvious theft” or “they knew” add nothing to the evidence and can be quoted back at you if the log is ever produced. Do not edit past entries; add new, dated ones if something changes or turns out to be wrong. A log that shows its corrections openly is more credible than one that looks flawless.

Decide on a rhythm and keep to it. A monthly pass over your most exposed images, and a quicker check after any new publication or licence, is often enough to keep the log current. Note in the log when you searched and found nothing as well: a dated record that an image did not appear on a given site can be as useful as one showing that it did.

Store the log alongside your original files and your records of publication. How to document a single use in detail, and how to preserve source files and metadata, are explained in the neighbouring articles.

Section 10

Objections that dated records help answer

When contacted, users of an image often raise the same points. Several of them turn on facts that a monitoring record can establish calmly.

• “We found it on Google.” A search engine indexes images; it does not license them. The record often shows where else the image appeared, which helps explain how it reached the site, but that does not remove the need for permission.
• “It was on a free site.” Free-wallpaper and aggregator sites often carry images without any rights to them. A copy found there does not authorise further use, and the log may show that the free site itself took the image from elsewhere.
• “We removed it.” Removal ends the ongoing use and is noted in the Case File. The dated captures show the period during which the image was online, which removal does not erase.
• “The photo is old.” Copyright in a photograph generally lasts for decades, and what counts is when the use took place. The capture shows that date.
• “Our web designer chose it.” Who selected the image is a matter between the operator and its contractor. Responsibility depends on the jurisdiction, but the operator is generally the party that published it, and the capture shows the image on the operator’s site.
• “You have no registration.” For works first published outside the United States, registration is not required before suing under 17 U.S.C. § 411(a), and the Berne Convention excludes formalities. For US works, registration affects what can be claimed and when a suit can be filed; it does not decide whether a use was authorised.
• “It is fair use.” Fair use under 17 U.S.C. § 107 is decided case by case on four factors, and the context recorded in the capture (a commercial page, a decorative use, the full image) is part of that assessment. The question is fact-specific and is not settled in advance by either side.
• “Send me a court order.” The matter is at the amicable stage. If legal action becomes necessary, Atlantic organises the file and may transmit it to qualified counsel in the relevant jurisdiction. Atlantic is not a law firm and does not give legal advice.

References17 U.S.C. §§ 411(a), 107 · Berne Convention

Section 11

Starting with Atlantic

To have your images monitored and your findings reviewed, submit them through the Atlantic platform. Atlantic verifies your authority to act and issues an authority certificate with a public verification address that recipients can check.

Occurrences are then captured with their fingerprints, reviewed for signs of a licence and, where the review supports it, opened as a Case File. Atlantic does not publish anything about the people it contacts, does not contact their clients, and keeps settlement discussions confidential.