A Recipe For Success?
AI has yet to replace the creativity of a real-life chef, but it does have other applications for luxury gastronomy, Professor Charles Spence argues.
There has long been interest in the question of how digital technology might transform our gastronomic experiences. However, when I first published Gastrophysics: The New Science of Eating back in 2017, it was impossible to foresee the incredible developments that have taken place in the world of artificial intelligence (AI) over the past few years.
AI, defined as human-like intelligence in machine form, comes in many forms, from service robots to conversational agents, and from machine vision to generative AI (GenAI). What does this mean for dining? At first sight, perhaps very little. Robot servers may have novelty value when they are first introduced, but their cost (both to purchase and to maintain) means that they have little practical value for either mainstream restaurants or fast-food chains.
As for high-end dining, many consumers have reservations about the integration of AI into gastronomy. One considerable challenge is that luxury is very much about the personal, the artisanal and the handmade. While AI may be able to deliver as far as personalised dish or drink recommendations are concerned, even the best chatbots fail to deliver the personal touch, not to mention the personal recognition, that many luxury diners expect. One of the most interesting areas for luxury gastronomy, however, relates to innovations in the area of GenAI.
The GenAI market has grown rapidly in recent years, and is projected to increase from $11 billion in 2020 to nearly $130 billion this year.
By 2032, the market is predicted to pass the $1.3 trillion mark. There are several potential applications for GenAI in the context of gastronomy. These span all the way from suggesting innovative combinations of ingredients through to the generation of novel culinary concepts, and from facilitating the development of immersive multi-sensory dining experiences through to offering exclusive personalisation in drinking and dining. In other words, GenAI has the potential to fuse data science and culinary artistry.
As such, GenAI is likely to play an increasingly important role in the luxury gastronomy of the future. That said, it will probably be encountered in the design of increasingly popular multi-sensory immersive dining experiences rather than recipes combining unusual combinations of ingredients.
Here at Oxford University, Giovanbattista Califano, Tianyi Zhang and I have been conducting research to assess people’s trust in GenAI recipes for both conventional and innovative versions of food and drink. In one study, a group of 201 people were asked whether they would be happy to trust AI (versus a cookbook) when searching for a recipe to make lasagne bolognese or a Negroni cocktail. While they were reasonably happy to rely on AI when it came to searching out recipes for traditional versions of food and drink, they were much less willing to trust technology when it came to innovative dishes such as tomato jam bacon lasagne or a white Negroni.
Such scepticism is probably merited given the negative consumer response to many of the innovative dishes and drinks that were recommended by IBM’s Chef Watson a few years ago. The problem when it comes to generating new recipes is that it is currently impossible to predict how a particular combination of ingredients will taste or whether people will actually like the result.
Take food waste, for example. The last few years have seen the emergence of a number of apps that promise to use AI to reduce waste by helping consumers create dishes from whatever ingredients they happen to have left in the fridge. However, enthusiasm for such digital solutions to the problem of food waste have been dampened by scare stories in the media where these apps recommended a recipe to create inedible concoctions.
As in so many other areas, GenAI tends to work best when combined with human creativity.
One particularly unappetising case from New Zealand involved a recipe that would produce chlorine gas, which can cause lung damage or death. As such, recipe generation by AI should be considered as a creative suggestion engine, but no guarantee of a tasty solution. The lack of the human touch when tech takes over is one of the dangers that those involved in luxury gastronomy should be aware of.
One of the most intriguing uses of GenAI in the luxury gastronomy sector is restaurant Krasota in Dubai (from $325 a head) where AI is used to deliver its immersive multi-sensory dining experience. In this case, GenAI has primarily been used to help enhance the quality and immersive-ness of the audiovisual storytelling elements that accompany the various courses on the tasting menu, rather than in the creation of recipes for the guests. For instance, an AI-generated version of French chef Paul Bocuse, who died in 2018, talks diners through the inspiration of his signature dishes such as quail and foie gras with truffle sauce that guests are about to eat.
Similarly, restaurants like Zenon, also in Dubai, now use AI-generated art installations to change the mood of the dining space. The Greek mythology-themed restaurant features dancing holograms and DJs wearing headgear that transforms their own movements into digital art projected onto the walls.
Over in London, chef Jozef Youssef of Kitchen Theory develops dining concepts for events such as the party held in Bush Creek, Wyoming back in 2023 for those who had paid the deposit for a flight on Virgin Galactic. The chef used Midjourney (a popular form of AI image-generation software) to create hyper-realistic images of a selection of culinary concepts. Once the client decided which dishes they liked the look of, the chef reverse engineered the ingredients and gastronomic approach needed to deliver the concept.
“We use generative AI platforms as tools for expanding our gastronomic creativity and pushing the boundaries of what we think is achievable,” Youssef says. “From unique flavour pairings and ingredient combinations, to original recipes, visual presentation and menu writing, these technologies offer a whole new creative instrument that we use to create even more immersive and multi-sensory experiences in real life.”
Such an innovative approach to the use of GenAI speeds up the process of culinary innovation while, at the same time, lowering concept development costs. It is, however, not without its drawbacks. While consumers are finding it increasingly difficult to distinguish GenAI gastroporn from human-generated content, once they learn that a food image has been generated by AI, they tend to like the food less. As such, it remains uncertain whether luxury dining experiences would be enhanced by the diner’s awareness of who (or what) is responsible for their gastronomic experience.
There can be little doubt that AI is going to play an increasingly important role in so many aspects of our daily lives in the years ahead, and the world of fine dining is no exception. However, because AI struggles to understand or predict the complexities of flavour perception, and thus is unable to reliably predict how much we will enjoy the experience of eating and drinking, my belief is that the role of AI in fine dining will primarily be focused on enhancing the overall experience, rather than coming up with the next flavour sensations.
Charles Spence is Professor of Experimental Psychology and head of the Crossmodal Research Laboratory at the University of Oxford. Connect with him on his LinkedIn profile.
Continue Reading
Why AI and Blockchain Will Decide the Future of Luxury
Redefining the Luxury Consumer Experience
Why Human Connection Reigns Supreme in the Age of AI Travel