AI menus say more about attitudes to AI than they do about the food

Recent examples have seen consumers describe AI food photography as uncanny, misleading and unappetising. The Guardian recently reported on restaurants facing significant backlash after customers identified AI-generated imagery in their menus and signage. [theguardian.com]

But is using AI to create cheap menu imagery actually that different from what the value end of the restaurant market has always done?

Budget restaurants, takeaways, food trucks and QSRs have rarely commissioned the kind of painstaking, expensive food photography we associate with restaurant brands that have deep pockets

They've historically relied on stock photography, food replicas, inexpensive graphic design and heavily retouched images. The objective isn't artistic authenticity. It's to communicate quickly and cheaply: here's roughly what we sell, here's what it costs, does it make you hungry? Generative AI feels like a logical evolution of that model.

For an independent takeaway or value focused QSR operating on thin margins, paying a photographer or designer every time the menu changes makes little economic sense. AI potentially offers a much faster, cheaper content-production model. So why the backlash?

I suspect we're looking at two different effects being conflated.

The first is execution.

Some AI-generated food simply looks terrible. Strange textures, impossible ingredients and overly polished surfaces can create an uncanny effect. That's a quality problem, and one that you also see with food displays that replicate the various dishes on the menu may want to order. Because they don’t look real, they feel inauthentic. This is a particular problem with food that has strong associations with freshness

But the second issue is more interesting. For some consumers, knowing that something is AI-generated changes how they evaluate it.

The image isn't necessarily being judged solely as an image anymore. It becomes a signal for cheapness, laziness, inauthenticity or even deception. Consumers are currently conflicted about AI. Whilst they can see enormous potential in the benefits AI could bring, there is also significant distrust and concern about its potential impact on society as a whole.  In one of Bryter's surveys of consumer attitudes to AI the most frequently cited concern about AI is that it can be used for misinformation. Simply using AI can trigger a lot of negative associations in consumers’ minds, and that matters even when it’s used for something as simple as a menu or display in a café.

So before using AI for this purpose the question shouldn't simply be “Can AI produce this content cheaply enough?”, it should also be "How will our particular customers interpret the fact that we've used AI?"

For a £6 meal deal in a takeaway, customers may be perfectly comfortable with inexpensive promotional production using AI. For an artisan bakery charging £8 for a sandwich, the same visual shortcut may conflict with everything the brand is trying to communicate about craft, provenance and authenticity. AI hasn't invented cheap content. It has made cheap content visible in a new way that some consumers currently find culturally uncomfortable.