← All guides Buying and selling

How to check if a product photo is real before you buy

Generated product shots sell items nobody made. Here is the four-minute check that catches them, in the order that finds fraud fastest.

· 7 min read · Best AI Image Detector

Reverse image search first, because a stolen photo is more common than a generated one. Then check the pixels. Then read the listing for the tells that a real seller does not have.

Why generated product photos work so well

A furniture listing needs one clean shot of a sofa in a bright room. That is the easiest thing a generator makes: no faces, no text, no hands, controlled lighting. Every visual tell people are taught to look for is absent from the photograph a scammer needs.

The economics also favour it. A seller can produce forty variations of a product that does not exist in a minute, list them across several marketplaces, and take deposits before a single buyer arrives to collect.

The check, in order

  1. Reverse image search the main photo

    Drop the picture into Google Images or TinEye. If it appears on a manufacturer's site, a stock library or another seller's older listing, the photo is stolen and the listing is not what it claims. This catches most fraud on its own.

  2. Run a pixel check on the same photo

    A generated image will not appear anywhere, because it did not exist until last week. That absence is not proof of anything by itself, so score the pixels. Check the image here — it runs in your browser and nothing is uploaded.

  3. Read the region map, not just the number

    A high score across the whole frame points to a fully generated product shot. One hot region in an otherwise clean photo usually means a real photograph with the logo, label or price tag replaced.

  4. Ask for a photo you specify

    Request a picture of the item next to today's newspaper, or with a specific object beside it. A generator can produce that too, but a scammer working at volume rarely bothers, and the delay itself is informative.

  5. Check the seller, not just the picture

    Account age, other listings, whether the same photograph appears under different names, and whether the price is plausible. Fraud is usually visible in the account before it is visible in the pixels.

91 94 89 96 93 90 88 95 92

All nine tiles above the decision line. Compare this with a real photo that has one edited area, where eight tiles stay cold.

A generated product shot scores high across the whole frame. There is no clean region because no part of the picture came from a camera.

What a fake listing looks like beyond the photo

  • One photograph only, or several shots from the same angle. A real seller has a phone and takes six.
  • No wear. Second-hand items have scuffs, dust and a room around them. Generated ones are showroom clean.
  • Backgrounds that repeat across supposedly unrelated listings from the same seller.
  • Price below the market by enough to create urgency, with a reason attached: moving abroad, house clearance, quick sale.
  • Payment off-platform. Bank transfer, gift cards or a deposit to hold the item. This is the actual crime; the photograph is the setup.
  • A new account with no history, or an old account whose earlier listings sold something entirely different.

Where checking is worth the four minutes

Not every purchase justifies the effort
SituationWorth checking?Why
Buying second-hand furniture or electronics from a strangerYesHigh value, off-platform payment, no buyer protection
A marketplace listing you will collect and inspect in personRarelyYou will see the item before money moves
A supplier sample photo before a bulk orderYesThe whole order rests on that one image
A rental property listingYesDeposits are paid before viewing more often than people admit
A retail purchase from an established shop with card protectionNoThe payment method already protects you

If you have already paid

Save the evidence before the listing disappears, because it will. Screenshot the listing, the seller profile and the messages. Download the original photograph rather than a screenshot of it. Export the check result so the score, the region map and the date are recorded together.

Then report it to the platform under its own fraud or synthetic media policy, and to your bank or card issuer. A card payment or a platform's protected checkout can often be reversed. A bank transfer usually cannot, which is why scammers ask for one.

What a generated photo cannot fake

A generator produces a picture, not a history. That gap is where most fraud is caught, and none of it requires looking at pixels.

  • A photo taken on request. Ask for the item beside something you name. A scammer at volume will stall, refuse, or send another stock-looking shot.
  • A video walkaround. Ten seconds of the item turning is still difficult to fake convincingly and almost never offered by a fraudulent seller.
  • Consistent surroundings. Real items sit in a real room. Ask for a wider shot and see whether the background matches the other photos.
  • A serial number or label. Legitimate sellers can photograph one. Generated labels fall apart under zoom, which is the one visual tell that has held.

Each of these costs the seller a minute and costs you nothing. Together they filter out almost every listing built on a generated photograph, before you have run a single check. A seller who will not spend that minute has told you something.

Questions people ask

Can you spot a fake product photo just by looking?
Sometimes. Look for lighting that is too even, an absence of wear on a supposedly used item, and backgrounds that repeat across listings. Current generators handle simple product shots well enough that a careful fake will pass a visual check, which is why reverse search and a pixel check matter more.
What if the image is not found by reverse image search?
That means it is not a stolen photo. It does not mean the item is real. A generated image is new and will not match anything, so an empty reverse search is exactly what both a genuine seller's own photo and a generated fake look like. Score the pixels next.
Do marketplaces check listing photos themselves?
Some run automated checks on volume sellers, and none catch everything. Enforcement is usually reactive, which means a listing is removed after buyers report it rather than before money moves. Treat platform vetting as a backstop, not a guarantee.
Is a high score enough to report a seller?
It is enough to report, and not enough to accuse. Platforms want the listing URL, the images and a description of what looks wrong. Include the score and the region map as supporting detail alongside the seller's behaviour, which is usually the stronger part of the report.