Consumer Trust Decides Which Parts Of A Beauty Campaign AI Can Touch

Beauty brands have multiplied the digital assets behind a single global launch, spreading them across e-commerce, paid social, retailer pages, email and packaging that changes by market. Creative teams have turned to AI to keep pace with the volume. The category has drawn a sharper line than most industries around where the tool can go, holding models, faces, before-and-afters, and anything supporting a product claim in human hands. Consumer trust and legal exposure set that boundary, and the strongest uses of AI in beauty happen before a campaign reaches an audience.

Elizabeth Schmidt is Executive Design Director at LLuMina Studios, an AI creative studio she founded for beauty brands. Schmidt led design for L'Oréal across its Luxe, Active and Consumer Products divisions, then ran an agency serving L'Oréal brands. Her work has covered Lancôme, Armani Beauty, SkinCeuticals, Kiehl's and CeraVe, and she recently spoke on artificial intelligence and creativity at Beautyworld Middle East in Dubai. Schmidt now splits her time between traditional creative direction and training beauty teams on where AI belongs.

"It's not ethical to use AI everywhere when you're selling skincare or beauty products that are promising a result. If you start seeing your trusted skincare brand using AI models, all that trust that's been built is going to collapse pretty quickly," says Schmidt. Through her work, she teaches creative teams inside global beauty houses where that line falls. The answer changes depending on which side of a photo shoot the work sits, and on which market the asset is built for.

Keeping up with volume

Finance departments and senior management push most creative teams toward AI, and the request usually arrives as a savings target. Schmidt points to a different driver inside beauty. Consumers moved beauty buying online after 2020, and brands answered by building far more for every launch while compressing the timelines. "Back in 2015 or 2017, when we would do a digital launch, we would need 40 assets for a new product launch for social," Schmidt says. "Now, if I do a global product launch for a brand, it's over two to three thousand assets, because you need e-com, paid, the website, e-retailers, Amazon, emails. Different sizes, different specs, and different countries require different labels."

A wording change from legal is what makes thousands of assets expensive. One revised claim propagates through the entire set, and Schmidt counts beauty brands now putting out eight to ten products a year where they once managed three or four. The teams absorbing that increase are the same size they were, and she describes widespread burnout among the creative directors and art directors she works with. "Imagine you build a whole campaign around 3,000 assets, and legal says this one claim has changed," Schmidt explains. "Now someone has to go and update 3,000 assets and save them at a certain size. That's why I started to look into AI. It wasn't let's save money or let's create new things with AI. It was how can AI help us keep up with this ever-growing need for online content."

Models, faces, and claims

Models, user-generated content, and before-and-after sequences are where Schmidt draws the hardest line. The teams she trains have asked her to leave the technique for building an AI model out of the curriculum entirely, so that a designer can't produce one by accident. "The human in the advertisements has to stay human," says Schmidt. "Anytime you feature a model, or UGC, or someone using it, or anything with skin, face, the actual human itself needs to stay human. Those are an area with beauty brands we don't even touch."

Product claims change what kind of image can sit beside them. Efficacy claims in beauty carry substantiation requirements, and a generated image next to a measurable promise puts the illustration out of step with the evidence behind it. Schmidt applies the rule to the whole asset, including the product shot next to the claim. "If I'm claiming that this moisturizer reduces wrinkles by 15 percent in six weeks, I can't use an AI image with that claim," Schmidt notes. "Legally we're claiming something, and then we're using generated images to support that claim. Anything that has a claim on it needs to be original photography, original models."

US shoppers are among the least receptive to synthetic imagery in beauty and skincare of any retail category. Other markets put the line somewhere else. Schmidt trains teams across several regions, which leaves the same asset facing different expectations depending on where it runs. She recalls observing those differences at Beautyworld Middle East, the region's largest beauty trade fair. "A lot of the brands were actually using AI models for their stuff at a trade show, which I was surprised about," Schmidt adds. "Maybe in the Middle Eastern market it wasn't such a taboo. But in the US, the consumers really don't want that, and they've been very clear about it."

Where AI earns its place

Schmidt puts AI to work on the product side of a campaign, in retouching, packaging fixes, and background and texture work. She also uses it for salvage jobs, on shoots that came back missing a market's pack, or carrying a label that has since changed. The largest gain she reports comes earlier, before a crew is ever booked. "Another place I'm really teaching the teams is in pre-production for a photo shoot," Schmidt explains. "Before you even go into your photo shoot, you're leveraging AI to show your concept, so that everyone is pre-aligned, from management to the photographer to everyone on set. You get a better output."

Internal presentations are the other place Schmidt points to. Beauty creatives spend a large share of their effort selling concepts up the chain before a product exists, and a generated visual moves that conversation faster than a mood board. Schmidt is direct with clients about what the tool will and will not return, particularly the ones arriving with a savings figure already attached. "My screen time is not down at all, if anything sometimes it's up," Schmidt concludes. "It's not that AI saves you all this time as a creative. It just allows you to do things you couldn't do before a lot easier."