The fashion industry is seeing a striking shift in how product photography gets made, now that traditional photo shoots are no longer the fixed, unavoidable cost they once were for every brand selling online. Daniel Baker, Head of E-commerce and Marketplaces at the British label Blue Vanilla, shared his hands-on experience with AI product photography in The AI Report 2026, published by fashion-tech platform The Interline (18 September 2026). He has been testing the technology since 2024, when his first AI-generated images carried every telltale sign of immature AI: poreless skin, fabric with no real drape, and vacant expressions his teenage daughter spotted at a glance.

The Cost Gap: From £2,500 to £50

The biggest difference Baker documents is the cost gap between the two approaches. A traditional shoot for fifty products – covering a model’s fee, a makeup artist, and a photography studio – runs to roughly £2,500 in the UK, or about £50 per product. With AI, the cost for the same fifty products drops to around £50 in total – just £1 per image. After Blue Vanilla built its own in-house tool in partnership with ViZO Studio, built on Google’s Gemini Image model (also known as Nano Banana), their actual cost fell further still, to around 15 pence per image.

The workflow itself stays fairly simple: photograph the product on a mannequin under ordinary lighting, upload it to the tool, describe the scene and the model you want, then choose the background and accessories. But Baker is clear that this apparent speed runs into a consistency problem – results can carry shadow errors or a hairstyle that shifts from one image to the next for what’s meant to be the same virtual model, which means manual review and extra touch-up work in software like Photoshop.

The Risks: Jobs, the Environment, and Creativity

At the same time, Baker doesn’t brush aside the ethical questions that come with this shift. Models, makeup artists and photographers stand to be affected the most, since their fees are likely to come under pressure to compete with AI’s near-zero cost – potentially making modelling less economically viable as a career over the long run. On the environmental side, Baker questions whether the energy and water that AI data centres consume genuinely offset the carbon emissions of traditional photo-shoot travel – and suspects they don’t. He adds a third dimension around creativity itself: if these models are trained only on past data, can they really produce anything genuinely new? These concerns sit alongside other unresolved questions about who owns a digital model’s likeness, and who holds the rights to clone it.

This tension between efficiency and ethical concern showed up clearly in H&M’s experiment building AI “digital twins” of thirty of its real models – with their consent and for a fee – to use in marketing campaigns instead of live shoots wherever possible. Even though the company stressed its commitment to protecting the models’ rights, the initiative drew significant pushback from industry workers who warned about the knock-on effect on photographers’, stylists’ and make-up artists’ jobs. Other major brands – including Zara, Zalando, Mango and Levi’s – have already begun using AI-generated model imagery at production scale, a sign that this shift is no longer a limited experiment but an accelerating trend across the sector.

What’s In It for the Next Generation of Fashion Designers?

• Understanding the basics of generative AI photography lets a new designer produce professional images for her own collection without needing a large photography budget – opening the door to launching a personal brand at a fraction of what it used to cost.
• The professional roles tied to photography are changing, not disappearing; Baker suggests photographers and stylists move toward new paths such as “prompt engineer” or “digital art director” – skills worth building into any design student’s development plan today.
• Awareness of the ethical questions surrounding the technology, from model rights to its environmental footprint, is becoming a core part of any emerging fashion brand’s decision-making – not a minor technical detail on the side.

The landscape, as Baker himself puts it, remains a mix of huge opportunity and risks that need managing carefully – and his advice to every brand is to experiment and learn fast, because whatever isn’t learned today could become tomorrow’s competitive edge in a rival’s hands.

Source: Daniel Baker, “Testing The Pace Of AI Progress”, The Interline, 18 September 2026.