AI product photography
What is AI product photography?
AI product photography is product imagery in which the scene, the background or the light around the goods is generated, while the product itself stays a real photograph. The tool drops a shot item into a new setting — a kitchen counter, a table, a street — or rebuilds the light on it. The goods are never invented: an image showing a colour, a material or a part the product does not have is a misleading advertisement.
Example
A home fragrance shop has 40 items, all photographed on white. The campaign needs them in a setting: a candle on a bedside table, a diffuser in a bathroom, a gift set on a dining table. A shoot with props means a day of shooting, hired set dressing and a second day of retouching.
Instead the real frame of each product is cut out and placed into a generated scene, with the light and the shadow matched to that scene. Across 40 items that is three scenes each — 120 frames in about a day and a half. The main image in the shop still stays the clean frame on white: marketplaces require a real photograph of the goods there, with no props and no graphics.
Why it matters for a business
The gain is pace and variety. A summer scene, a Christmas scene and a bathroom scene come out of one frame, so the advertising changes without a new shoot. For a small catalogue that is the difference between two campaigns a year and one a month.
The risk sits in accuracy and in the rules. The image must not mislead about the main characteristics of the goods; EU consumer law treats that as an unfair commercial practice, and the parcels come back as soon as the customer unwraps something else. Since 2 August 2026 the transparency rules of the AI Act, Regulation (EU) 2024/1689, apply as well: the tool marks its output in a machine-readable form, and you disclose the image as generated when it could pass for a photograph of a real scene.
What to ask
- Which frame stays real? The product is photographed; only the scene around it is generated, never the goods.
- Who checks colour and material against a sample before the image reaches the shop?
- What rights does the tool grant over its output, and may the image run in paid advertising and in print?
- How is generated content disclosed, and does the marking survive retouching and upload?
- Is there a face in the frame? A real person needs consent, and a generated face must not resemble a specific one.
- Who keeps the source files and the prompts, so the same scene can be rebuilt six months later?
We photograph the real products as part of photo and video production; cutting out, matching colour and preparing the variants for the shop and for advertising are part of photo retouching.