AI photos on Allegro starting August 2, 2026. What must the seller label?
AI Photos on Allegro starting August 2, 2026. What must the seller mark?
Starting August 2, 2026, sellers on Allegro will have to pay closer attention to how they prepare product photos. The new rules will cover entirely generated visualizations, realistic backgrounds created by AI, and virtual models presenting products, among other things.
For sellers marking listings, this means they will need to check how the materials they use were created. This matters not only for compliance with the platform's rules but also for buyer trust and sales effectiveness.
AI Photos on Allegro – what changes on August 2, 2026?
The changes announced by Allegro are related to the entry into force of further provisions of the European Union's AI Act, which will take effect on August 2, 2026. The regulation introduces an obligation to inform recipients about images generated or modified by artificial intelligence if they can create the impression of being authentic. The goal of the new regulations is not to ban sellers from using AI in listings, but to increase transparency. The buyer should know whether they are looking at a real photo of the product or an image whose significant elements have been created or altered using artificial intelligence.
A separate issue is the way in which Allegro intends to help sellers fulfill this obligation. The platform announced the release of a beta version of a tool for marking AI photos by August 2, 2026, at the latest.
How will the marking of AI photos work?
Allegro plans to provide sellers with a beta version of a tool for marking photos generated or modified using artificial intelligence. When adding a photo, the user will be able to indicate that AI was used to create it. This option is expected to be available in several listing publishing and updating channels:
Product Catalog,
listing creation form,
process of adding listings via file,
API used by external systems and integrations.
After selecting the appropriate option, Allegro will automatically place a visible "AI" watermark on the photo. The seller is to receive a preview of the image before saving it, allowing them to check the appearance of the mark and how the photo is presented in the listing.
However, this does not mean that the platform will independently recognize and classify all materials. The responsibility for indicating photos that used artificial intelligence remains with the seller. The person publishing the listing should know whether the image was generated from scratch or if artificial intelligence created its key elements, such as the background, person, or product variant.
With sellers managing larger numbers of listings in mind, Ofertator is preparing a solution designed to make it easier to mark photos created or modified using AI. The feature is intended to allow for quick identification of such materials, even in listings prepared beforehand, without the need to manually analyze and edit each product individually. Organizing photos in advance will make adapting listings much more efficient.
Which product photos may require marking?
Not every photo prepared using modern graphic tools will be subject to the same rules. What matters most is whether the image elements have been generated or significantly modified by artificial intelligence in a way that buyers might mistake for a real photograph of the product. Allegro encourages helping buyers verify the authenticity of photos.
Product generated from scratch
Marking may be required for an image showing a product that was not photographed but created entirely using artificial intelligence. This applies, among other things, to visualizations of custom-made furniture, decor designs, garden structures, or personalized accessories.
A similar situation occurs in the case of prototypes and visualizations prepared before the start of production. Even if the final product is supposed to look exactly like the graphic, the buyer should know they are viewing a generated image and not a photo of a finished unit.
Real product on an AI-generated background
Marking may also be needed when the product itself is real, but artificial intelligence created the surrounding scene. An example would be a photo of a chair placed in a generated living room or a tool presented in a fictional workshop. Such a graphic can look as if it were taken in a real location, which is why the recipient should receive clear information about how it was created.
Virtual model presenting the product
Another category consists of photos showing virtual people. If AI has generated a model presenting clothes, shoes, a handbag, a watch, or jewelry, the image may require marking. This also applies to situations where a real photo of the product is digitally overlaid onto a generated figure.
Similar rules may cover the presentation of cosmetics on a virtual face or hand, as well as interior design visualizations where AI has created realistic residents. Marking does not take away from the value of such a presentation, but it allows the buyer to distinguish generated material from a real photo shoot.
Does every photo edit using AI require marking?
Simply using a program equipped with artificial intelligence features does not automatically mean that a photo should receive the "AI" tag. What matters most is the scope of the change made and its impact on how the buyer perceives the product. If the tool was used solely to improve photo quality, crop it, or remove minor technical flaws, the result can still represent the real product without creating a fictional situation.
Modifications that generate elements not present when the photo was taken or that clearly change the appearance of the item must be assessed differently. If the generated scene looks like a real photo shoot, the customer may mistake it for an authentic representation of the product. The obligation to mark may therefore be determined by the final result, and not just the name of the program or the feature used.
Usually does not require marking:
cropping a photo,
adjusting brightness and contrast,
basic color correction,
retouching technical flaws,
preparing a clean, white background without creating a fictional scene.
May require marking:
generating a realistic environment,
generating objects that were not in the photo,
changing the appearance of the product,
creating a non-existent variant,
using a virtual character,
extending larger parts of the image with a generative feature.
Does the term "deepfake" also apply to product photos?
Commonly, deepfake is associated primarily with videos in which the face or voice of a specific person has been swapped using artificial intelligence. However, this understanding is much narrower than the definition adopted in EU regulations.
According to the AI Act, a deepfake can be an image, audio recording, or video generated or modified by AI. The content must resemble existing persons, objects, places, entities, or events in a way that could falsely induce the recipient to believe it is authentic or true.
This means that the concept of deepfake is not limited to human faces. It can also include products, furnishings, buildings, gardens, or entire arrangements if the generated image looks like a real photo. An example would be a real piece of furniture placed in a non-existent living room or a product presented in an environment created by artificial intelligence.
What about photos added before August 2, 2026?
Initially, Allegro's tool is intended to allow marking only for newly added photos. This means that sellers will not be able to immediately mark the use of AI in every photo already in an active listing. However, the platform does not rule out extending this feature to older materials in the future. Therefore, it should not be assumed that photos published before August 2, 2026, will be permanently excluded from the marking system.
The second issue is the AI Act itself. Article 50(4) requires disclosing that deepfake content has been generated or modified by AI, but it does not contain a general exemption for images created before these regulations started to apply. In practice, not only the date the file was created but also its continued use in the listing may be relevant. Since the interpretation of obligations regarding older photos may still evolve, the safest solution is an early audit of materials instead of waiting for Allegro's tool to expand.
A seller can prepare for the changes by:
making a list of listings that contain visualizations,
identifying photos generated or modified by AI,
contacting manufacturers and wholesalers regarding the origin of the photos,
recording information about the program and method of preparing each file.
Sellers using Ofertator will be able to use the upcoming feature to automatically flag every AI photo in their managed listings. This solution will reduce manual work, especially with an extensive product catalog.
Will Allegro automatically check all photos?
Allegro announces that in the future it may use metadata stored in files to recognize materials created or modified using artificial intelligence. However, this does not mean an automatic check of every photograph published in listings. The platform clearly indicates that the responsibility for correctly assessing and marking the image still lies with the seller.
August 2, 2026: Allegro is to make the beta version of the tool available at the latest, allowing sellers to declare AI use and add an automatic watermark to photos.
December 2, 2026: from this date, AI tool providers are to place hidden information about the use of artificial intelligence in files. Allegro notes that it may use this metadata in the future but does not declare widespread, automatic checking of all photos.
Seller's responsibility: the person listing the offer must determine if the photo requires marking. This also applies to materials received from a manufacturer, wholesaler, photographer, or agency.
How can the AI mark affect sales?
Possible drop in trust
Some buyers may view the "AI" mark as a sign that the photo does not represent the actual product. This is particularly true for listings where appearance, color, texture, or proportions are highly significant in making a purchasing decision.
Possible increase in credibility
However, openly marking materials can also work in the seller's favor. Transparent information about the use of artificial intelligence shows that the listing is not trying to mislead the buyer. For some customers, this way of communicating will be proof of honesty and a professional approach to presenting the product.
Product-image match is what matters most
The "AI" mark itself does not have to automatically lower conversion rates. A much larger problem can arise when a generated image presents a product more attractively than it looks in reality. If a customer receives an item that differs in color, finish, size, or details from what they saw in the listing, the risk of returns, complaints, and negative feedback increases.
Therefore, the seller should treat AI as a tool supporting presentation, not as a way to beautify the product beyond its actual features.
Frequently Asked Questions from the Allegro Community
Does a visualization made in Photoshop require marking?
This depends on how the image was prepared. Manually cutting out a product and placing it on an existing background does not automatically mean AI was used. However, if the seller used Photoshop's generative features to create a realistic environment, modify the product, or add new elements, the photo may require marking.
Does a generated background with a real product need to be marked?
Such a photo may be subject to marking, even if the product itself was not modified. The buyer should know that the presented interior, garden, terrace, or studio is not the actual place where the photo was taken, but a scene created by artificial intelligence.
Does changing a product's color using AI require marking?
If AI has created a color variant that was not actually photographed, the safest solution will be to mark it. This also applies to situations where the product in that color is actually on sale, but its photo was created by digitally altering another variant.
Is a mention of AI use in the listing description enough?
A note in the description alone will be insufficient, as the customer often sees the photo before reading the content of the listing.
Can I add my own AI mark?
Your own information can increase transparency, but it will not replace the solution provided by Allegro. It is safest to use the mechanism prepared by the platform, especially when it becomes linked to the listing form, catalog, and API.
What are the penalties for incorrect marking or lack of marking?
The consequences will depend on the circumstances, platform rules, and the nature of the material. Failing to provide the required information can lead to intervention by Allegro, the need to correct listings, reports from other sellers, or loss of customer trust.
Summary
The changes regarding AI photos on Allegro will not ban the use of artificial intelligence in online sales. Their goal is primarily greater transparency and clearly informing buyers about how the image was created. Starting August 2, 2026, sellers will be able to use Allegro's tool to mark photos generated or significantly modified by AI. It is also worth remembering that the responsibility for assessing the photo will remain with the seller, even when the materials come from a manufacturer, wholesaler, or agency.
Before the changes take effect, it is worth:
reviewing active listings and used photos,
determining which images were created using AI,
checking the origin of photos from external suppliers,
keeping information on how the files were prepared,
developing a simple procedure for marking new materials,
combining AI visualizations with real product photos.