AI-Generated Images in E-commerce: What Does the Law Say? (EU AI Act, 2026)

From 2 August 2026, the EU AI Act (Article 50) requires transparency for certain AI-generated or manipulated content — notably deepfakes and content likely to mislead the public. For a product photo that faithfully shows a real garment on a synthetic model, the main point of attention is depicting a person. The safest practice: use clearly synthetic models and add a transparency notice.

This article summarizes what the law requires, what concretely affects fashion brands, and the best practices to put in place now. (This content is informational and not legal advice; check your situation with a lawyer.)

Do you need to disclose an AI-generated product image?

The question comes up constantly: "Do I have to state that a photo on my product page is AI-generated?" The answer depends on what the image shows.

• If the image faithfully reproduces a real product (the garment the customer will receive), it doesn't mislead about the merchandise.
• The sensitive subject is depicting a person: an AI model must not imply that a real, identifiable person wears or endorses the product without their agreement.

Hence a simple rule: synthetic models + transparency remove most of the risk.

What does the EU AI Act (Article 50) say?

The EU AI Act is Europe's framework for regulating artificial intelligence. Its Article 50 covers transparency and provides, among other things, that:

• AI-generated or manipulated content (images, audio, video) that may constitute a deepfake must be labeled as such;
• the information must be given in a clear and recognizable way.

Key timeline:

Beyond the European text, consumer protection rules (banning deceptive practices) and image rights already apply in most countries.

What this concretely changes for a fashion brand

For an e-commerce business using AI visuals, the practical implications are:

1. Favor synthetic models. A generated face matching no real person avoids consent and image-rights questions.
2. Stay faithful to the product. The garment shown must match what's delivered: color, cut, material. It's also the best way to limit returns.
3. Plan a transparency notice. Anticipate displaying an "AI-generated visual" indication where context requires it.
4. Document your pipeline. Know which visuals are generated and how, to answer any request.

This is exactly the approach of a virtual try-on studio like ClothesLook, built around synthetic models intended for commercial use, with no impersonation of real identities — a principle detailed in our article on faceswap and ethics.

How to display an "AI-generated image" notice

Several approaches, often combined, enable clear transparency:

• a readable text notice near the visual or in the product description ("AI-generated visual");
• a metadata marking in the image file;
• a discreet watermark for the most sensitive content.

The goal isn't to devalue the visual, but for the buyer to understand its nature unambiguously — which builds trust rather than diminishing it.

Transparency: an asset, not a constraint

Far from a hindrance, transparency about AI use can become a brand argument: it signals modern, controlled and responsible production. Consumers increasingly value clarity; displaying it proactively sets you apart.

In summary

From 2 August 2026, Article 50 of the EU AI Act requires labeling certain AI content. For a fashion brand, the safe path is to use synthetic models, stay faithful to the product, and plan a transparency notice. Done well, this transparency strengthens trust.

To produce visuals aligned with these principles, explore ClothesLook or read our guide on diverse AI fashion models.