I want to make one thing clear right off the top: This column is not going to attempt to defend AI slop on restaurant menus. That would be impossible, because some of it is truly indefensible. See the latest viral exhibit below.
The death of folk graphic design in favor of AI slop across our country's delis is leading to some incredibly Lovecraftian food horrors pic.twitter.com/snDyC675Ei
There is simply no world in which that “shrimp” should have seen the light of day.
Now, this is an egregious example of a practice that is now quite common. And I’d argue that as AI becomes more deeply ingrained in everyday life, restaurants deserve a little more grace for taking part in it, even if the results aren’t always pretty.
Restaurants, especially small ones, often have good reasons to use AI to create things like menu images or social media posts. These operators are typically strapped for time and capital. They almost certainly don’t have a marketing department. AI makes these tasks easier.
“What it does do is it saves us time, really, and resources to try and compete in a marketplace, especially in an industry that, you know, [a large percentage] of diners in New York City are dead or dying,” restaurateur Jordan Beau told The New York Groove in April.
Beau’s restaurant, Daisy’s Diner in Brooklyn, has come under fire for using AI to generate much of its social media content. Its TikTok account is full of too-slick-to-be-true videos promoting purple ube pancakes and what it says is Brooklyn’s best happy hour.
@daisys.diner.and.bar.nyc Looking for the perfect spot to kick off your evening? Stop by Daisy’s Diner and Bar in Park Slope for the best happy hour in NYC! From our legendary cocktails to our signature farm-to-table, scratch-made comfort food—including juicy sliders, creamy mac n’ cheese, and our house-made bar mix—we’ve got everything you need to start your night right. See you at the bar! #DaisysDinerAndBar#FYP#NYCHappyHour#VIRAL#ScratchKitchen♬ original sound - Daisys Diner & Bar
It’s all a bit hard to square with the restaurant’s classic diner vibe and farm-to-table menu. One could argue that Daisy’s would be better off doing nothing on TikTok than spamming its followers with AI. And yet, in terms of getting more eyeballs on Daisy’s brand, the strategy has worked: Almost every video has tens of thousands of views.
While some restaurants have eagerly embraced the power of generative AI, others have done so reluctantly. Robert Trakolfer, co-owner of Zenith in Pittsburgh, began using it to create an image of the restaurant’s weekly menu for Facebook and Instagram because his usual editing program was no longer supported.
This sparked an outcry from followers who complained that it was hypocritical of Zenith to use AI while claiming to support local artists. (In addition to vegetarian food, Zenith also sells vintage goods and art and has hosted live music and poetry in the past).
The backlash had an impact on Zenith’s business, Trakolfer said. But he argued that the anger was misplaced. As a poet and writer himself, he is wary of AI, and his support of the arts over the years is well-documented. He turned to AI in a pinch, he said, and has since found a workaround. But he believes dunking on restaurants for using AI misses the forest for the trees.
“The thing that bothers me the most is AI is incorporated in everything at this point,” he said, including cellphones, photo editing software, and Facebook, the very platform people were using to attack his AI menus. “Picking out a few little businesses that use it because they can't afford to hire an artist to render their imagery is pointless.”
Not only that, but AI for years has been foisted upon restaurants as the answer to their every problem, even as the technology itself has proven to be hit or miss. Now they’ve found a useful application for it, only to be told that they’re doing it wrong. We’re in the early stages of an unprecedented technological revolution, for heaven’s sake. Everyone, in some sense, is still figuring it out.
And while restaurants’ predilection for AI slop is often framed as some kind of moral failing, a lot of it boils down to aesthetic differences. AI-generated images understandably make many people cringe, but there are plenty of others who don’t seem bothered by it and may even find it appealing somehow (circular shrimp aside).
This is part of life. I can’t stand the look of Marvel movies, and I’ll never understand why some drivers choose to clutter their dashboards with rubber duckies. But both are entrenched in the mainstream, so who am I to judge?
It should also be noted that people have a major incentive to gripe about AI-generated abominations: It’s low-hanging fruit that tends to get a ton of engagement on social media. Nevermind that it often comes at the expense of a well-intentioned, if misguided independent business — clout-chasing takes no prisoners.
Now, I say all that, and yet … I just can’t shake the image of those giant, Cheerios-shaped shrimp. Who could have possibly thought that was a good idea? I’m still trying to get an answer to that question.
Trakolfer of Zenith, who is generally sympathetic toward restaurants’ use of AI in this way, was also stumped.
“I can understand people just not being able to afford to pay somebody to create images for them, but it's easy to take photographs and post that,” he said. “But, you know, whatever they want to do, that's their business.”
Indeed it is, and the market will ultimately decide its fate.
I’d rather not entertain one possible explanation for the proliferation of AI nightmare food: That the restaurants creating it simply don’t care. And if that’s the case, they probably would not have lasted long anyway.