Can we use AI to make before-and-after photos for skincare ads?

Not to show results. An image that shows an effect the product did not produce, or exaggerates one it did, is misleading under advertising law in most markets. AI can illustrate how a product works in a clearly non-photographic style, beside genuine and documented photographs.

Can genuine before-and-after photos be retouched at all?

Only in ways that cannot affect the claim, such as removing dust from the lens or matching the overall crop of the two images. Anything touching the area where the product works, including skin, hair, teeth or a stained surface, risks misleading viewers and should be left alone.

Can renovation companies show AI images of a finished project?

Yes, as long as they are presented as visualisations of a proposed design rather than photographs of completed work. Label them clearly, and use real photographs when the image is meant to prove the quality of work the company has already done.

AI Before-and-After Images: The Rules

Before-and-after imagery and AI.

A before-and-after image is evidence of a result, so AI should almost never make one. Generating or enhancing the "after" to show what a product does is misleading advertising in most markets, and in cosmetics, weight loss and health it is a well-worn route to regulator rulings. AI is safe for clearly illustrative visuals and for laying out genuine, documented photographs, never for creating the result.

Why a generated after misleads

Viewers read a before-and-after pair as proof: this is what happened to a real person, room or surface. An AI-made after shows a result nobody achieved, and an AI-enhanced one shows more than the product delivered. Either way the image itself is the claim, and an unsupported claim. The UK Advertising Standards Authority made the principle explicit in 2021, when it ruled against influencer ads on social media for tanning products whose images had been filtered, finding that filters exaggerating a product's effect are misleading. Generative edits are the same problem with more power.

Where the rules bite hardest

Cosmetics and skincare, weight loss, hair and teeth treatments, aesthetic procedures, cleaning products and home improvement all lean on before-and-after imagery, and all attract complaints. In the EU, cosmetic claims must meet the common criteria of Regulation (EU) No 655/2013, including truthfulness, evidential support and honesty, and the Unfair Commercial Practices Directive (2005/29/EC) covers misleading commercial practices towards consumers generally. In the US, the FTC requires claims to be substantiated, and its Endorsement Guides treat a testimonial result as typical unless the ad clearly discloses what people can generally expect. Medical and weight-loss advertising carries extra sector rules in most markets.

What AI can do safely

Illustrate a mechanism in a style that nobody could mistake for a photograph: a drawn cross-section of skin, an animated diagram of how a cleaner lifts grease. Lay out and size genuine before-and-after photographs consistently across formats, without touching the subject. Show a proposed renovation or interior as a clearly labelled visualisation of a design rather than as finished work. Produce everything around the evidence, such as backgrounds, typography and the rest of the campaign, while the evidence itself stays a real, unaltered photograph.

Documenting genuine results

A genuine pair should be shot under the same light, from the same angle and distance, with the same styling and make-up, so the difference visible is the product's effect and nothing else. Keep the raw files, dates and the person's consent. Make sure the result is typical, or say plainly what results people can generally expect. Keep a record showing the photographs were not generated or altered; in Synthetic White, every asset carries a record of which model made what, which helps keep generated material apart from untouched photographs.

Updated 26 October 2026