When Levi’s announced in 2024 that it would begin testing AI-generated models to “supplement” its human casting pool, the backlash was swift and the conversation it sparked has yet to settle. Nearly two years later, fashion brands are still wrestling with a question that has no tidy answer: is it ethical to replace human models with synthetic ones, and if so, under what conditions? As generative AI tools become cheap enough for a solo Etsy seller to use on a Tuesday afternoon, the stakes have expanded well beyond luxury marketing departments and into the entire retail ecosystem.
The Economics Driving the Shift
The appeal is, at its core, financial. A single professional photoshoot with a model, photographer, stylist, and studio rental can run anywhere from $5,000 to $50,000 depending on scale, and that’s before usage rights and licensing fees are factored in, which can add another 20-40% to the total spend. AI-generated imagery, by contrast, can produce a usable product image in minutes for a fraction of the cost. For small and mid-sized brands operating on razor-thin margins, this isn’t a marginal efficiency gain — it’s the difference between running a seasonal campaign at all or skipping it entirely.
This is precisely why tools built for smaller sellers have proliferated so quickly. Print-on-demand shops and independent apparel brands, who could never have afforded traditional model photography, now have access to PixelPanda’s free AI t-shirt mockup generator with real-looking models, letting a one-person Etsy operation present a hoodie or graphic tee on a photorealistic model without hiring anyone or scheduling a shoot. Used transparently, this kind of tool democratizes access to professional-grade visuals rather than displacing an existing workforce — an important distinction regulators and industry bodies are increasingly trying to draw.
Where the Ethical Lines Actually Fall
The controversy isn’t really about AI imagery existing — it’s about disclosure, displacement, and diversity. Three issues keep surfacing in industry discourse:
- Disclosure: Should consumers be told when a “model” in an ad isn’t a real person? The UK’s Advertising Standards Authority has already signaled that undisclosed synthetic models could run afoul of misleading advertising rules, and France passed legislation in 2023 requiring labels on retouched or synthetic commercial images.
- Job displacement: Model booking agencies report that entry-level and mid-tier catalog work — historically an entry point for new talent — has softened noticeably as brands experiment with AI for basic product shots.
- Representation: Critics argue that AI models, trained on datasets skewed toward conventional beauty standards, risk reversing hard-won progress on size, race, age, and disability representation in fashion media, unless brands deliberately intervene in how these systems are trained and prompted.
How Brands Are Responding
Some companies have taken a middle path. H&M’s 2023 announcement that it would create “digital twins” of real, consenting models — paying them for the rights to their AI-generated likeness — was widely seen as a template for ethical use, because it kept human beings, and their compensation, in the loop. Mango’s AI-generated Teen campaign drew more criticism, largely because it appeared alongside real photography without clear labeling, blurring the line for consumers.
The practical guidance emerging from these case studies is fairly consistent: label AI-generated imagery clearly, compensate any real person whose likeness informed the model, and avoid using synthetic imagery in campaigns explicitly built around diversity or body positivity messaging, where authenticity is the entire point. Brand safety, in other words, now includes AI disclosure the same way it once had to absorb influencer disclosure rules.
The SEO and Discovery Layer
There’s a less-discussed dimension to this shift: AI-generated product imagery still has to get discovered, indexed, and ranked like any other content. Brands investing in synthetic model photography for product pages should pair that visual strategy with basic technical SEO hygiene — proper alt text, structured data, and optimized meta descriptions — so search engines can accurately surface the products being shown. A free meta tag generator for collection pages can help smaller retailers ensure their AI-generated visuals are actually supporting discoverability rather than existing in a vacuum, since flashy imagery means little if the page never surfaces in search results.
As Green Stock Media has reported in its coverage of synthetic media across advertising verticals, the imagery question is increasingly inseparable from the trust question — consumers are becoming more sophisticated at spotting AI content, and brands that get caught obscuring its use tend to pay a reputational price that outweighs whatever production costs they saved.
What Comes Next
Regulation is likely to tighten before it loosens. The EU’s AI Act, along with state-level transparency laws in the US, are expected to impose disclosure requirements on synthetic media in commercial contexts within the next two to three years. Brands that build labeling and consent practices into their workflows now — rather than retrofitting them under legal pressure later — will be better positioned when those rules land.
The technology itself isn’t the ethical problem; how transparently and equitably it’s deployed is. Fashion brands that treat AI models as a tool for expanding access rather than quietly cutting costs at the expense of human talent will likely find themselves on the right side of both consumer sentiment and the coming regulatory wave.