When Saving Money Starts to Show

I have started noticing something while walking around Paris.

AI-generated images are appearing everywhere. I’m not talking about the spectacular examples that circulate online. I mean the much more ordinary ones I now see in the Métro, on posters, restaurant menus and advertising panels. At first I barely noticed them. Now I find them difficult not to see.

There is often something familiar about them. People look attractive in a strangely interchangeable way. Sometimes there is an obvious mistake in the background. Increasingly, there isn't.

Advertising has never been a particularly reliable representation of reality. Photographs have been staged, retouched and manipulated for as long as the industry has existed, and not every AI-generated image is bad. Some are technically convincing. What bothers me sits somewhere else. So many of these images feel cheap.

The images are often trying very hard to look expensive: beautiful people, beautiful interiors, perfect sunsets, luxurious materials, everything polished. And yet after seeing enough of them, I have the opposite reaction. The image doesn't make the product feel more valuable. It makes the communication around it feel less so.

A conventional advertising photograph can involve a photographer, models or actors, assistants, stylists, hair and make-up artists, producers, retouchers, set designers and locations. Depending on the campaign, many more people may be involved. Generative AI removes most of that cost. From a company's point of view, the attraction is obvious: an image that once required a production can now be made far more quickly and cheaply. There is another cost, though, that doesn't fit in a spreadsheet.

In February 2026, Gucci released a set of AI-generated images as part of its Primavera campaign. The visuals, including a woman in a fur coat in a restaurant, a pair of legs emerging from a car and two models against the night sky, were mixed into a rollout that also included traditional photography, and each AI image carried the caption "Created with AI." There was no attempt to mislead anyone. The reaction online was still striking: "cheap," "tacky," "AI slop."

I don't find the episode interesting because Gucci should or shouldn't use AI. Brands experiment, and some experiments provoke reactions nobody fully anticipated.

What interests me is the reaction itself. With Gucci, craftsmanship, history and creative direction are already part of what people are paying for. The images around the brand are judged in that context too. When the imagery associated with such a brand suddenly looks inexpensive to produce, the saving becomes visible.

Not the AI, necessarily. The saving.

The same image in the window of a small neighbourhood restaurant reads differently. A shop with no budget for a photographer and a full crew now has access to imagery it could never previously afford, and that access is a genuine gain. The calculation becomes more uncomfortable when the company behind the image has no such limitation, particularly when craftsmanship and creative work are part of the value it sells.

Paris has survived plenty of bad advertising before, and generative AI won't ruin it either. But some brands may be discovering an uncomfortable consequence of making their imagery dramatically cheaper to produce.

Consumers may notice the saving too. And once they do, the image may end up costing the brand more than it saved.

***

The photographs are gathered here:
Selected Works