Everyone’s trialling AI-generated product photography to try and cut costs, produce faster results, and increase output. Some of it is really impressive. And some of it looks like physics took the day off!
As many of you are discovering, the quality of your output depends almost entirely on the quality of your input. And your input requires the combined knowledge of art direction, photography and graphic design.
If you want a beautiful directional light that makes your packaging glow, then you’ll need to understand lighting. Want a depth-of-field that draws the eye to your hero product? You need to understand optics. Want water droplets that look like actual water being poured or splashed, and not like water that’s trying too hard. You need to understand how water moves. AI doesn’t. It just makes it look exciting.
Garbage in, garbage out. It just looks shinier now.
There’s also a problem nobody’s talking about yet: the tools have a house style. AI image generators have aesthetic tendencies – the glowing orb, the beaming light, the hero product sitting in a liquid with a ripple effect. If you’re using the same tools with the same default outputs as every other small beauty brand, you’re not building a brand world. You’re sharing one.
And your customers are noticing, even when they can’t exactly articulate why. Research from the Nuremberg Institute for Market Decisions found that when consumers were shown identical ads, one labelled AI-generated and the other labelled human-made, they rated the AI version as less natural, less useful, and were less willing to purchase. Same image. Different label. Fewer sales. In skincare, where the entire purchase decision is built on trust, that’s not a risk worth taking.
The platforms are already responding. Amazon now requires sellers to disclose when product images have been generated or substantially enhanced by AI, with misleading imagery resulting in listing penalties regardless of how they were created. Etsy goes further, mandating explicit disclosure in every listing description. The platforms aren’t banning AI imagery. They’re drawing a line at deception, which tells you everything about where buyer confidence actually sits.
Here’s where it gets concerning.
According to Getty Images’ Building Trust in the Age of AI report (which surveyed over 30,000 adults across 25 countries), 98% of consumers agree that authentic images and videos are pivotal in establishing trust. And nearly 90% want to know if an image was created by AI. That’s not a fringe concern. That’s almost everyone who follows you.
And if your brand skews younger, pay attention to this. An IAB study found that four in five advertising industry executives believed consumers feel positive about AI-generated ads. Fewer than half of Gen Z and Millennials actually did, with Gen Z having the strongest negative reaction. There is a significant gap between what brand decision-makers believe and what their customers actually feel. If you’re in the room where the “let’s use AI for content” decision gets made, this is the number you need on the table.
Your customer can’t always tell you why they didn’t buy. They just scrolled past.
If you’re an Australian brand, this hits even closer to home. A joint study by the University of Melbourne and KPMG (surveying 48,340 people across 47 countries) found that while half of Australians use AI regularly, only 36% are willing to trust it, and 78% are concerned about negative outcomes. Australia ranks among the lowest globally on acceptance, excitement and optimism about AI, and only 30% of Australians believe the benefits outweigh the risks – the lowest of any country surveyed. Your Australian audience is already primed to question what they’re seeing. Giving them a reason to doubt your imagery is the last thing your brand needs.
It’s worth remembering that we’ve seen this pattern before. Gartner calls it the Hype Cycle: a new technology arrives, excitement peaks, and businesses rush to adopt it before they fully understand how to use it effectively. We’re sitting right at that peak with AI image generation. The slide down the other side is where brands that moved too fast start to feel it.
When Canva arrived, everyone believed they could do their own graphic design. The tool was accessible. The skill behind using it well was not. The result was a wave of content that looked like it came from a tool rather than a brand. AI image generation is following the same pattern – but in product marketing, the stakes are higher. Your product photography is often the first impression a customer has of your brand.
What skincare photography actually needs to communicate
For skincare brands specifically, the stakes are even higher. Your customer is making a decision based on how they think the product will work for their skin. That decision is almost entirely visual. They’re reading texture, finish, and quality from your imagery before they read a single word of copy.
AI struggles to realistically reproduce what skincare photography actually needs to communicate:
-
- Texture and finish. The difference between a lightweight gel, a rich cream, and a facial oil is communicated almost entirely through how light reacts with the product. AI renders most skincare textures as a generic glossy or matte surface. It doesn’t understand viscosity or how a serum catches light differently from a cream.
- Colour accuracy. Skincare formulations have very specific colours – a vitamin C serum has a particular amber tone, a retinol product has a specific opacity. AI generates plausible colours rather than accurate ones. For a brand where the formulation is part of the story, that matters.
- Ingredient authenticity. If your brand uses real botanical ingredients as a visual prop, AI will produce something that looks like those ingredients from a distance but won’t hold up to scrutiny.
- Packaging detail. Skincare packaging is often where brand investment is highest – embossed logos, frosted glass, metallic pumps, dropper bottles. AI renders these as a general impression rather than a specific object. Labels become decorative rather than legible.
If you need to write pages of prompts and codes to explain your brand, your products, your positioning, and what you want to create, at what point is it actually faster or cheaper than working with a human who already understands all of that? A human who knows your category, can see what your competitors are doing, understand your customer, and has sat in on enough conversations to know what challenges your brand is facing. That knowledge doesn’t need prompting. It shows up in the work.
Great product photography isn’t about getting things to look pretty. It’s about making your product feel true to your brand in a way that stops a specific customer mid-scroll. That requires decisions – about light, angle, surface, colour, texture, story – that no default output can make for you.
The brands that get this right won’t be the ones who used AI the most – or the least. They’ll be the ones who stop to ask: Does this align with our brand values, our customers’ expectations, and the quality standards our product deserves?
Want to learn more?
Curious about what’s possible for your brand? Get in touch, I’d love to talk!