On Instagram, the models showing off Liz Muthoni’s thrifted dresses have flawless skin, confident poses, and chic backdrops. But there’s a twist: none of them exist.
Instead of hiring models, photographers, or renting locations, Liz uses artificial intelligence (AI) to bring her thrift business to life. She uploads a photo of the clothing, often on a mannequin, then prompts AI to create a model. She’ll specify size, list the bust measurements, add ethnicity, and even describe the background she wants. Other times, she feeds the software a Pinterest image for inspiration. Within minutes, she has polished images ready for her online shop.
“It saves me time, money and stress,” Liz says. “Before, I’d spend hours arranging shoots, paying models and photographers, and scouting locations. Now, I’ve cut almost three-quarters of my costs. It’s just me, my phone, and AI.”
Liz Muthoni of Lizzie Fashions also known as Thrift Loft at her stall on August 20, 2025.
Photo credit: Francis Nderitu | Nation Media Group
Her story is part of a growing wave. Across Kenya, young entrepreneurs are experimenting with AI to keep their businesses afloat. Nowhere is this more visible than in fashion, where image is everything and AI promises cheap, sleek images at the click of a button.
Liz’s discovery of this use of AI was accidental. She had started using ChatGPT to design posters and cards for her shop—Thrift Loft. One day, she wondered, could the same tools dress up her mannequins?
“I realised I don’t need to exhaust a model, making her change outfits ten times in the sun. I don’t need to pay a photographer every weekend. And I don’t need to rent spaces. AI gives me all that,” she says.
On her Instagram page, the results look like any other fashion feed: women posing in sundresses on leafy boulevards, chic tops styled in urban coffee shops, flowing gowns set against a sunset on a beach. Only Liz knows that behind each picture is a mannequin in her tiny store.
Liz Muthoni of Lizzie Fashions also known as Thrift Loft taking photos of some of the outfits on sale at her stall on August 20, 2025.
Photo credit: Francis Nderitu | Nation Media Group
For customers scrolling online, the scene is already set, the vibe is curated, and the backdrop is polished, making it easier for them to be drawn to the clothes. But what Liz celebrates as liberation is also a disruption. Established designers are watching with both curiosity and caution.
Ogake Mosomi, a leading bridal designer in Nairobi, admits AI is opening new doors.
“It’s definitely making presentation more affordable to people who don’t have the budget,” she says. “For showcasing ideas, it does a pretty good job.”
Yet, she quickly adds, AI has limitations, especially in high-end fashion.
“The patterns it shows aren’t always accurate. The fabric textures, the cuts, sometimes they look perfect, but they can’t exist in real life. Clients come with these AI-generated images asking me to replicate them. Some of those designs are impossible to make,” she says, laughing softly. “AI has increased people’s expectations. They want 3D gowns at bargain prices. They think anything is possible.”
For her, AI is a tool useful in sketches and mock-ups, but not a replacement.
“I’d use it alongside hand-drawn sketches or 3D mini-conventions. But fashion is human. Seeing a real person in a gown feels different from a flawless computer image.”
While AI may be a threat in the fashion industry, some are choosing to adapt. Smith, a Nairobi-based photographer, is surprisingly hopeful.
“For me, I’m embracing it,” he says. “I even pay for ChatGPT to generate images for mood boards.”
He explains how he now uses AI to develop concepts before shoots: “I can type out an idea, get a visual sketch, then build a real shoot around it. It makes me more creative.”
“It won’t take away my job. It assists me. You just have to move with the change,” he says.
Smith envisions a hybrid future, where AI assists creatives instead of replacing them.
Still, he admits, it may reduce demand for services.
“Yes it has cut down on jobs. But on the bigger shoots, weddings, high fashion, AI can’t replace the human touch. At least, not yet.”
That “yet” haunts many in the fashion industry. Around the world, fears are mounting that AI could wipe out jobs for models, photographers, stylists and even designers.
In Kenya, where the creative sector already struggles with unemployment, the worry is sharper. With AI evolving, models may not be needed anymore. That means fewer gigs for young women who once relied on Instagram shoots for pocket money. It means fewer weekend bookings for freelance photographers.
But at the same time, AI has given Liz and many like her a survival edge. Running a thrift business is competitive, and margins are thin. By slashing her costs, Liz can keep prices low and attract more customers.
Liz Muthoni of Lizzie Fashions also known as Thrift Loft taking photos of some of the outfits on sale at her stall on August 20, 2025.
Photo credit: Francis Nderitu | Nation Media Group
Beyond economics, there’s a cultural question: whose beauty is AI projecting? Ogake is sceptical.
“The bodies AI generates are not realistic; exaggerated hourglass figures with tiny waists and oversized hips; the proportions simply aren’t right,” she says.
Many AI models default to Eurocentric standards. Unless carefully prompted, AI often struggles to capture African beauty in all its diversity.
“Sometimes I tell it I want a curvier model, or a darker skin tone,” Liz says. “But even then, the results can look off. It takes trial and error.”
The risk, experts warn, is subtle. If AI keeps pushing unrealistic images, consumers may internalise skewed beauty ideals. What happens to body positivity when all the “models” are algorithmically flawless?
Around the world, fashion houses are testing AI models. Levi’s announced a partnership with a digital model agency to “increase diversity” in campaigns, though critics blasted the move as fake diversity. Luxury brand Balmain used AI to create a virtual influencer.
What Liz is doing from her small Nairobi apartment is, in many ways, part of a global experiment. Only hers is born of necessity, not novelty.
“The way AI shows fabric is not always accurate. A lace dress in the image may not look like the real thing,” she says.
Liz is aware of the risk. She often clarifies to customers that the clothes they see are the real ones, just displayed differently. “I have to explain to clients that the clothes are thrift”
For Ogake, the way forward is clear: “It’s a long-term opportunity, yes. But I don’t think AI will replace traditional fashion. Real-life garments, human models ... those can’t be erased.”
