Ask Runable forDesign-Driven General AI AgentTry Runable For Free
Runable
Back to Blog
Technology7 min read

Ugly AI Food Photos Are Only the Beginning | WIRED

Restaurant menus with AI-generated food images are going viral for their off-putting aesthetic. But AI’s impact on modern menus runs even deeper. Discover insig

aiartificial intelligenceimage recognitionfoodkitchen+2 more
Ugly AI Food Photos Are Only the Beginning | WIRED
Listen to Article
0:00
0:00
0:00

Ugly AI Food Photos Are Only the Beginning | WIRED

Overview

I can’t stop staring at the AI-slop images of Mexican food I see during my morning walk to work. On the menu posted outside, the taco truck displays lunch options that appear uncanny and unappetizing, as if an alien who’s never eaten human food before drew a taco from memory. This food truck is not alone in putting AI-generated images on its menu. It’s everywhere.

AI menu slop is now going viral for being disgusting, and these posters seem to be an evolution of the “Chat GPT flyer pandemic.” Whether it’s a maggot-infested burrito or a plasticky pasta, consumers are expressing online how gross these fabricated images look. The aesthetic is distinctively off-putting, sometimes with random holes throughout a dish or otherworldly cuts of meat served as an entree.

Details

The images are immediately noticeable but only scratch the surface of how generative AI tools have already overhauled restaurant menus. Sellers aren’t just generating AI images of burgers, they’re also relying on AI to decide what goes in between those two buns.

The birth of the online menu set us on the path toward these monstrosities. More consumers started deciding what to eat based on online menus and by ordering delivery from their phones to their doorstep. Faced with growing pressure, restaurant owners digitized their menus and created additional visuals.

Stores wanted to capitalize on delivery sales and leaned into generative AI images, vying for a visibility boost on third-party apps like Door Dash without paying for a photoshoot. While the most powerful models currently on the market can output photorealistic images of food, many of these images appear to be generated by tools with worse capabilities, like the default setting of a free chatbot.

“I went to order pizza one night on Uber Eats,” says Brendan Sweeney, CEO and cofounder of Popmenu, a digital marketing software company for restaurants. “There were five different pizza places that were using the same AI-generated image.” He says if the food image appears “completely inauthentic” and “obviously AI,” then the restaurant is making a worse decision than foregoing photography altogether.

What was originally generated to satisfy digital demand is now creeping its way into the physical world. AI menu slop expanded from its online origins into posters in coffee shops and signs displayed as deli menus. Restaurant owners sometimes seem oblivious to how off-putting low-quality image generations look.

Noticeably AI-generated images stand out on modern menus for all the wrong reasons. But AI tools are influencing what appears on your local coffee shop’s menu, beyond just images.

Tastewise, one of the largest data insights companies in the food and beverage industry, partners with companies like Nestlé, Pepsi Co, and Kraft Heinz. It’s currently all-in on “agentic” automations for every step of the food creation process.

Tastewise’s library of automations is meant to help companies decide what foods to sell and how to sell them. There’s an agent to help you track how product launches are going, an agent to test new concepts (banana creatine smoothies, anyone?), and an agent to find opportunity gaps in current menus.

“I think the bigger question for them is how do they guide and train the AI on what should be in that burger and what should be on the menu,” says Michael Bortinger, Tastewise’s director of marketing and partnerships. “We have billions of data points understanding what consumers are doing with food, what’s actually happening on menus and in the retail landscape.”

The result is Taste GPT, a chatbot-style tool powered by food-specific insights. Companies can pay for access to the bot, though the website offers you a few free queries, so I tried it out.

Two of the suggested prompts were about what competitors are doing with protein-enhanced foods and the current trending flavors in plant-based snacks. When I picked the latter, Taste GPT flagged dates as a “standout rising star” and “cake-flavored plant-based snacks” as another relevant—though not as hot—trend. Looking at what’s already available, I’m not sure I’d ever buy some cake-drizzled mini-rice cakes.

Bortinger rattles off cold foam, hot honey, and strawberry matcha as flavors and menu additions that are high on the consumer interest list. “Ube is a trending flavor, but that also may be starting to be past its peak,” he says. When I asked about pickles, as a lover of the limited-time dill-flavored Grinch-themed fries from Mc Donald’s, he assured me that companies are still finding “new formats” for delivering the innovative, pickle-y goodness. The dill-pickle pizza slices I keep seeing on Instagram do look quite delicious.

The tension between food marketing and reality existed before the generative AI boom. Food stylists for commercials have long used creative methods to present objects of consumption, like swapping ice cream for similarly textured mashed potatoes or painting the perfect grill marks onto a burger patty. While those changes often make the food seem more tantalizing than reality, most of the cheap menu slop I see day-to-day still looks puke-inducing.

Despite their many mistakes in deciding how to present food images, restaurant owners will keep turning to AI tools to automate aspects of their businesses and optimize their online offerings. Around 85 percent of restaurateurs currently "feel comfortable" using AI tools and plan to use them even more extensively in the future, according to a recent poll from Toast.

Rather than using fully AI-generated images, Sweeney wants restaurants to use software—like Popmenu—to enhance images of real food captured by the smartphones in their pocket. “AI can now very easily patch those up, plate them all the same, and make the lighting consistent,” he says. “I think that’s where the value is.” He sees these enhancements as a professional way to present the dish, based on what they are actually selling.

Not all food sellers want to put AI on their menu, though, and for good reason. These synthetic images of food are becoming a cultural touchstone in the broader, highly effective AI backlash. Popular local restaurants are going viral for their refusal to participate, posting handwritten notes about not using generative AI posters for promotion.

We are living under the uncanny reign of the AI slop menu. In a few years, even the cheapest AI image of food may look more tempting than these current iterations. That said, I don’t mind a slightly ugly, albeit real, photo of dishes on the menu when I dine out or order delivery. If anything, I prefer their more authentic rough edges.

When I walk home from work today, I think I’ll finally give that food truck a chance, keeping my fingers crossed that the tacos on my plate look better than the AI ones on the poster.

In your inbox: WIRED's most ambitious, future-defining stories

In your inbox: WIRED's most ambitious, future-defining stories

AT&T Top Deals: Wireless, Internet, Bundles and More!

Take Up to 25% Off Custom Prints with Vista Print Promo Code

Key Takeaways

  • I can’t stop staring at the AI-slop images of Mexican food I see during my morning walk to work
  • AI menu slop is now going viral for being disgusting, and these posters seem to be an evolution of the “Chat GPT flyer pandemic
  • The images are immediately noticeable but only scratch the surface of how generative AI tools have already overhauled restaurant menus
  • The birth of the online menu set us on the path toward these monstrosities
  • Stores wanted to capitalize on delivery sales and leaned into generative AI images, vying for a visibility boost on third-party apps like Door Dash without paying for a photoshoot

Cut Costs with Runable

Cost savings are based on average monthly price per user for each app.

Which apps do you use?

Apps to replace

ChatGPTChatGPT
$20 / month
LovableLovable
$25 / month
Gamma AIGamma AI
$25 / month
HiggsFieldHiggsField
$49 / month
Leonardo AILeonardo AI
$12 / month
TOTAL$131 / month

Runable price = $9 / month

Saves $122 / month

Runable can save upto $1464 per year compared to the non-enterprise price of your apps.