As generated imagery becomes normal in e-commerce, the question changes from "was there a camera?" to "is this the product?" Here is the standard and a checklist.

The short answer

A synthetic product image is authentic when it represents the real product faithfully: the same colour, fabric, texture, drape and proportions. Whether a camera was involved matters less to shoppers than whether what arrives matches what they saw. A 2026 conversation on The Interline framed the test as one question, whether the product is represented faithfully, and the industry is converging on that standard.

Key takeaways

  • Authenticity in product imagery now means faithfulness to the product, not the presence of a camera.
  • Faithful means colour, fabric, texture, drape and proportion match the real item.
  • Even strong systems produce images that fail checks; one vendor fixes about one in ten with a retouching team.
  • Build review into the workflow: check every image against the product before it is published.

What does "authentic" mean for an AI image?

On The Interline's podcast Is There Such A Thing As An Authentic Synthetic Image?, Eran Dagan of Botika reduced the question to one test: “is the product represented faithfully?” If colour, fabric, texture, drape and proportions are accurate, he argued, the customer is satisfied, whatever produced the image.

We think that is the right standard for product imagery, and a useful one because it can be checked. A lifestyle scene can be invented; the product inside it cannot. The background, the model and the light are creative choices. The garment, the bag or the shoe is a promise.

Do AI product images still need human checks?

Yes. Dagan described Botika's own process: most images pass the company's internal checks, but around 10% still need fixing, which it handles with a dedicated retouching team. Those are the vendor's own figures, but the shape of the process is the important part. Generation is fast; review is what makes it safe to publish.

Teams that skip review save a little time and risk a lot: a return because the colour was off, a complaint because the logo was distorted, or a quiet loss of trust in the brand's product pages.

What should you check on every product image?

Why does faithfulness matter commercially?

Because the image sets the expectation the product has to meet. The Interline's spotlight on fit and body data cited National Retail Federation and Statista figures estimating that returns cost US retail around $890 billion in 2024. Not all of that is about images, and fit is a bigger driver in apparel, but an image that misstates colour or proportion invites a return that a faithful one would avoid.

There is a second, slower cost. Shoppers learn which brands' pictures can be believed. Faithful imagery is a trust asset; unfaithful imagery spends it.

How should teams build review in?

  • Start from the real product. A photo of the sample or the approved render, not a description.
  • Keep the image tied to its style and colourway, so reviewers compare against the right reference.
  • Review before publishing, every time, with the checklist above.
  • Record the approval, so anyone can see who signed off which image.

Should brands say an image was made with AI?

Rules on disclosing AI-generated imagery differ between markets and are still changing, and some retail partners have their own requirements, so check both before you decide. Our view is that disclosure and faithfulness answer different questions. Disclosure tells shoppers how an image was made. Faithfulness tells them whether they can trust what they see.

A disclosed image that misrepresents the product still fails the shopper. An undisclosed image that represents it exactly may still breach a platform's rules. Brands need both: honesty about method where it is required or expected, and a product that matches the picture every time. Of the two, faithfulness is the one that decides whether the parcel comes back.

  • Keep a record of how each image was made and who approved it.
  • Follow each market's and each retailer's disclosure rules.
  • Never let disclosure stand in for accuracy.

What does this look like in practice?

An illustrative example. The brand and the details are invented to show the idea; they are not a case study.

A bag brand generates on-model and lifestyle images for forty SKUs from product photos. Most look right. In review, the team catches a few problems: a gold clasp rendered slightly brassy, a monogram slightly warped in one angle, and a strap drop that looks longer than the spec.

Because each image is tied to its style and colourway, the reviewer compares it against the approved swatch and the flat rather than against memory. The flagged images are regenerated or retouched before publication. The faster generation step did not make review optional; it made review the step that protects the brand.

Where CotonAI fits.

Product fidelity over pretty pictures is one of the bets CotonAI is built on. Product photography, AI fashion models and lifestyle scenes start from your real product and are built to keep shape, texture, labels and colour true to the original. Every image belongs to the right style and colourway, and every image can be reviewed and approved before it is published. See AI product photography.

Sources.

We read these pieces and wrote this analysis in our own words. Figures are credited to the source each article gives; vendor and company figures are their own claims.

  1. The Interline, Is There Such A Thing As An Authentic Synthetic Image?, Ben Hanson with Eran Dagan of Botika, 15 October 2026. Podcast with a vendor
  2. The Interline, Spotlight On: Fit & Body Data, 22 September 2026. Sponsor-supported instalment

Common questions

Is AI product photography authentic?

It is authentic when the product is represented faithfully: colour, fabric, texture, drape and proportions match the real item. Creative elements like the scene can be invented; the product should not be.

Do AI-generated product images need checking?

Yes. Even strong systems produce some images that fail checks. One vendor told The Interline it fixes around 10% with a retouching team.

Can AI product images reduce returns?

Faithful images set accurate expectations, which helps avoid returns caused by colour or proportion mismatches. Fit remains a larger driver of apparel returns.

What should I check in an AI product image?

Colour, material and texture, proportion, logos and labels, hardware and trims, and consistency across the set.

Keep reading

Show the real product. In every image. We'll show you the workflow end to end and recommend where to start. Book a demo