All around the world, we’re already eating like people of the future. To begin close to home in Dubai, at the tech-savvy pizza chain Dodo, chef Spartak Arutyunyan recently tasked ChatGPT to suggest a recipe that might appeal to customers of this city. The bot came up with a kofta-like topping in a sauce of tahini, sumac and za’atar, which proved tasty enough (with a few tweaks) to go on the menu.
Across town, the high-end Zenon Restaurant & Lounge now uses AI-powered Microsoft Kinect cameras to monitor the interior and feed those images back to the resident DJ, who then turns them into digital projections that interact with the music, enhancing the house aesthetic of ancient Greek myth meets hyper-modernity.
Meanwhile, more than 100 professional kitchens in the city have deployed systems like Winnow Vision to reduce wastage. Optical AI tracks each item that goes in the bin and makes corresponding adjustments to purchase orders and menu listings. So far, this is saving the domestic food service industry about AED 10 million per year.
This is only scratching the surface of the estimated 1.7 billion tonnes of food that go to waste worldwide (according to the UN’s Food and Agriculture Organization), but it’s a good start in addressing the problem. The same goes for every possible application of Artificial Intelligence to what and how we eat. As ubiquitous as it has come to feel lately, we’re still barely beginning to understand the potential of this tech.
Consumers might first need to distinguish the merely ‘futuristic’ from the real shape of things to come. So-called ‘smart’ restaurants of the sort that have recently opened from Tokyo to Barcelona – with sci-fi robot servers and tabletop digital tablets for fully automated ordering – are not necessarily attuned to popular demand.
“The opposite is proving to be true so far,” says Erich Eichsetter, a former head chef and designer who has come to specialise in food tech innovation at San Sebastian’s Digital Gastronomy Lab. In tracking worldwide trends, Erich has noted the relative failure of such business models, which don’t seem to grow beyond their initial novelty.
“These kinds of automated restaurants might be fun to go once, but people don’t really go back. If I was to speculate how the future will be, I can see this kind of tech being used to optimise human interaction, not to get rid of it. We might see a robot running back and forth from the kitchen, bringing the food, opening the bottles and so on, while waiters as we know them become something more like hosts, or entertainers, or conversators.”
Generally speaking, the work of the Digital Gastronomy Lab is geared toward the service side of the industry, developing AI-based products or projects designed to save users time or money, “which amount to the same thing”. One example would be a tool for optimising inventory systems, by which chefs could issue voice commands to a stock-taking database.
“So instead of writing ‘12 one-litre bottles of olive oil’, you can just say it, and it goes direct to the system.” Erich and his team won funding for their ‘proof of concept’ from the Basque Government, as part of an initiative to foster new ideas at the intersection of tech and food. The Basque Country is indeed home to one of the world’s great gastronomic cultures, but it does not automatically follow that creatives in the field are happy to embrace the latest labour-saving devices.
“In my experience, I’d say that managers are in favour, for sure. But operators, not so much. I think it has to do with chefs and cooks being artisans. They work with their hands, and they want to feel empowered that a workspace is theirs and they don’t like being told what to do in it.”
As to the culinary art itself, Erich does not buy the argument that the machines will eventually crunch enough data to learn everything there is to know about food, and replace human experts. “I always use the example of digital music,” he says. “There will always be place for orchestras playing classical pieces on traditional instruments. But in reality, most players today are electronic musicians, just because it’s much more scaleable. It doesn’t mean they are less creative, it’s just a different approach. And in food production, someone who uses hydroponics is not lesser than someone who grows the old-fashioned way. Creativity is ubiquitous. There’s no constraint on it.”
Some of the most groundbreaking AI applications are being tested at the production side of the industry, where machine-learning extends to so-called ‘biohacking’. Certain algorithms, spliced with computational biophysics, can now generate novel proteins and enzymes, target and manipulate health-promoting genomic properties, and tailor specific ingredients around intolerances or allergies. Thus, you get successful start-ups such as Redbloom making a fiery hot sauce that doesn’t trigger IBS symptoms, for example. Or the Chilean company NotCo, now a market leader in synthesising vegan substitutes for popular animal products. “They apply AI to formulate, say, a vegan milk, with the same proteins and flavour compounds but without the animal fats,” explains Erich. “At a domestic level, you could use AI in this way to formulate a bread, and make it more fluffy or more crusty just by changing the parameters.”
At a corporate level, meanwhile, the sheer operating speed of this tech can cut months or even years off the long, expensive process of research and development for new food products. Unilever created a low-salt bullion using AI to perform a complete flavour analysis in a matter of days, and Kraft applied Quantative Data Analysis to replicate an original tomato-based product with 94 per cent accuracy, factoring in sensory metrics like ‘mouthfeel’.
When it comes to the biggest challenges facing the food industry – namely, nourishing a planetary population of 8 billion and rising, even as climate change threatens crops and conflict disrupts supply chains – the hope must now be for these learning machines to ‘think’ equally big in terms of solutions. Erich says, “We are the most empowered generation in human history because we have the most information at our disposal. I lived in Bolivia for two years, and I met farmers using smartphones on remote mountains in the middle of nowhere. Whoever has the curiosity can plug into this information, and if knowledge is available to everyone, that might be only the thing that will help us.”
The information flows both ways, of course, and we are also the data. AI is already reading us by way of our social media accounts and online purchases, and facial recognition systems will soon be extracting our demographic details whenever we eat out. “Things like your age range, gender, whether you’re dining alone, at what time of day and so on,” says Erich. “These metrics can be incredibly effective for increasing sales.”
Tech that’s been standard for years in hotels and air travel will also dynamise restaurant pricing, so that popular items cost more at busy times, or dishes that are quicker and easier to prepare will appear first on digital menus.
The real trick, however, is what Erich calls “DNA personalisation”. “It’s still in early phase,” he says, but as an exercise in speculation his team has already published a report on The Restaurant of The Future. “We imagine a kind of food passport, with your information safely stored on it, and a system that processes this data to personalise recipes and so on.” Which is to say, a dining experience designed in real time according to your tastes and nutritional requirements. An experiment along these lines at the Harvard Innovation Lab presented a sensory survey instead of a menu, asking test “customers” about their preferences.
An AI algorithm then designed the menu accordingly and sent it to the lab kitchen. “The results showed that the survey group liked the food they were given, but they also tried things that they wouldn’t have otherwise.” This kind of interaction, says Erich, is likely to define your dining experience in years to come.
Eerie, presumptuous, or intrusive, perhaps. But also so finely calibrated to our wants and needs, our likes and dislikes, that we may well feel we’re being fully catered to for the first time in our lives.
