The sameness problem behind those unappetizing AI-generated menus
When it first happens to you, you think youâre crazy. You wander into a cafe and look at a menu with a variety of bagel sandwiches, but each illustration looks eerily flawless, precisely symmetrical, and oddly smooth, eliciting a visceral sensation that something isnât right. You might think youâre paranoid, but youâre not losing your mind. Generative AI menus have hit the restaurant business courtesy of models courtesy of models trained on a narrow, âpleasingâ aesthetic that produces a look that feels wrong even when you canât articulate why. Sometimes, these illustrations are egregiously fake, like a burrito with cheese so bubbly and melty that it looks more likeavant garde artthan lunch. More often, theyâre soordinarylookingthat you only notice something iswrongwhen you take a second to look more closely. âItâs almost like an alien trying to make a pizza without understanding its core principles,â Reality Defender CTO Alex Lisle told TechCrunch. (Reality Defender itself is part of a growing category of startups selling AI-detection and content-verification tools â a business that exists in part because of issues like this one.) Throw a rock in NYC and youâll hit a mildly off putting AI food advertisement.pic.twitter.com/lLXVkFIFoG Lisle says that the way these models are built can help explain why illustrations seem to embrace such a specific aesthetic â one where every ice cream scoop is perfectly round, and where shrimp seem to have been genetically modified to eat their own tails, creating new âLovecraftian food horrors.â Large language models (LLMs) and diffusion models â the kinds of AI models that make seemingly omniscient chatbots and image generators like ChatGPT and Midjourney possible â are trained on vast quantities of data. The models then identify patterns in the datasets to predict what a user is looking for when they ask something like, âMake me a menu for a burger restaurant.â âA lot of this stuff looks like a Chiliâs menu from 2015, and thereâs a reason for that,â Lisle said. âThat was the corpus of work from which [the models] drew their function.â New training data is invaluable to the companies building AI models â Amazon has even been found to source rare books to scan and add to its training data, only todestroy those booksonce theyâve been uploaded. Itâs inevitable that some AI-generated content will seep into these incomprehensibly large data sets. But when AI models train on too much of their own AI-generated content, they riskmodel collapse. âModel collapse is almost like a mad cow disease⊠when you feed the outputs from one model back into itself, eventually the inbreeding becomes too much, and the whole thing collapses,â Lisle explained. âWhat we see here is convergence, which isnât necessarily model collapse.â Convergence is a bit less extreme, degrading the quality of an AIâs outputs without making it entirely useless. If someone asks an AI model to generate a menu for a fast food restaurant, the model will likely reference menus from Wendyâs, Burger King, McDonaldâs, or another popular chain. These menus already share a similar style, which means that the AI-generated outputs will mimic that same style, only to further reinforce it further if the AI-generated menu ends up back in training data. But menus and advertisements for food will always look better than the real thing, like a Big Mac in a McDonaldâs commercial where each layer of the sandwich is arranged by a prop designer to look maximally appetizing. This effect can become even more pronounced in AI outputs. âThe optimization of the data sets is for pleasingness, or you know, not being offensive, and so thereâs a way that turns into homogenization,â Lee Rainie, Director of the Imagining the Digital Future Center at Elon University, told TechCrunch. âWhat AI is known to do both in images and language is to shave off the edges.â On a more localized scale, this smoothing of images seems to happen when you use an AI image generator to create a menu and apply edits to it. On X, a user namedLabtecshowed what happens when you make a menu in ChatGPT, then edit it 100 times to see how the food continues to look less and less like it should. (We replicated the experiment and found similar results.) âThe end result actually makes me uncomfortable,â Labtec wrote. I made a restaurant menu in ChatGPT then edited it 100 times to see how those hideous slop menus end up the way they are. The end result actually makes me uncomfortablepic.twitter.com/sDIUKlabdp Restaurants are likely falling victim to this problem, revising their AI-generated menus to alter small details over and over, like prices or item names. It seems that with each edit, the food images become a tiny bit more round and smooth. âPeople have an almost unexplainable sense about when theyâre looking at something thatâs AI-generated, compared with something that was real in the first place,â Rainie said. âThereâs just a sensibility that people sometimes find hard to articulate, but they kind of know it when they see it and I think thatâs one of the reasons why some of the early stories about the backlash [against restaurants using AI menus] is so pronounced.â Thereâs science behind our aversion to these AI menus. Researchers at the University of Duisburg-Essen in Germanyfound that AI-generated food images exhibited an âuncanny valleyâ effect, where images of food that looked almost real elicited more disgust and unease than images that were obviously fake. That squeamishness only intensifies in light of the cultural context around AI. If people react to these images so negatively, then thatâs probably reason enough for restaurants to stop trying to make AI menus work. But the issues that bring us perfectly browned hamburger buns extend beyond the dinner table. âSeeing and hearing has always been believing, to the point where even our court systems are entirely tuned to the idea that the gold standard in evidence is taped confessions and videotaped evidence,â Lisle said. âThatâs no longer the case. The world has fundamentally shifted, for good or for ill.â
