Body Technique vs Computational Gastronomy

An almost identical point to Chef Cracco’s had been made only a few weeks earlier by Ruby Islam, a famous chocolatier based in Hyderabad, central India. In late January, the local Literary Festival featured a ‘Science and the City’ event, with a session devoted to ‘The Future of Food in the Age of AI’. There, Islam also argued that AI can never substitute a human chef’s cooking touch. 

In doing so, she was seemingly relying on a core tenet of cultural anthropology. In the wake of Marcel Mauss’s cutting-edge investigations, anthropologists have long recognised the different manual acts involved in preparing and, hence, consuming food as techniques of the body—in Mauss’s own words, ‘the constant adaptation to a physical, mechanical, or chemical aim’ performed through one’s bodily gestures. 

Both Cracco and Islam might be assuming that AI will never be able to shake or even reach that bedrock of craftmanship. Perhaps they intuitively take comfort in Mauss’s idea that techniques of the body such as the ones they have mastered can remain themselves only insofar as they combine effectiveness and tradition—that is, an individual’s organic connection to the education he has received from his community, where he has a specific and recognisable role. 

Yet the anonymous, horizontal, and individualistic nature of AI, at least in the forms currently known, appears to by-pass that very connection. Ruby Islam’s conversation partner at ‘Science and the City’ made that challenge rather visible. The panel, in fact, also featured Prof. Ganesh Bagler, a trailblazing figure in the emerging field of computational gastronomy. Far from being steeped in the localised traditions of a certain creative mindset, Bagler’s work aims to establish how the molecules of ingredients can create flavours and successful combinations of taste without going through the time-consuming, subjective exercise of trial and error in flavour selection—a foundational process in the formation of culinary cultures and their distinctiveness. 

More broadly, the fact that shaping flavour can be construed as a process dependent on trial and error and molecule-identifying, rather than as a craft escaping exact categorisation, hints at the impact AI has already had in altering the discourse on food-making. That is also significantly vaster than amateur cooks, scrolling through videos of alluring recipes in their Instagram feeds, might have guessed. 

Meta, Systemic, and Global: A Three-pronged Revolution

The application of AI to the domain of food has already altered creative processes at a meta-level, from production to consumption. 

In his fifteen-year-old assessment of digital gastronomy and its advancements, Amit Raphael Zoran at the Hebrew University of Jerusalem described AI-empowered kitchens as ‘a critical side for reimagining global food systems’. That’s more ambitious than preparing a memorable roast in the here-and-now, no doubt. This AI-friendly strategy also underpins the UN’s attention to global nutrition, which incidentally earned the World Food Programme a Nobel Prize for Peace as early as in 2020. 

As Zoran observed, the fact that AI tools can easily scratch off personal and cultural considerations in shaping food-related suggestions makes them singularly suitable to assess abstract diets, implement them through food programmes, or establish resource-efficient food systems that prioritise environmental impact. Surely, AI can deliver highly efficient insights into substituting lamb-based dishes, with all their symbolic poignancy, with more ‘sustainable’ fibres in your Easter menu, or replace centuries of tribal bread-baking practices with less water-consuming techniques. 

Flavour pairing is another domain where the deployment of AI-related tools on culinary techniques is becoming prominent—and is shifting paradigms of creativity. Jun Pyo Seo at the University of Seoul has recently put forward FlavorDiffusion, a data-based, AI-operated framework designed ‘to predict food-chemical interactions and ingredient pairings’, with the specific intention of moving beyond the traditional expertise of human chefs. 

Another issue FlavorDiffusion aims to tackle is the confinement of certain pairings to specific cuisines, which in turn makes those very combinations hard to generalise. In other words, AI might be bringing back globalisation with a vengeance, at least in matters of food consumption. If that sounds far-fetched, the story of how celebrity company Palantir equipped international food chain Wendy’s with AI tools to optimise its infrastructure—from inventory to menu planning and reservations—confirms the suspicion: what AI does best, in terms of facilitating cooking skills, is reducing its complexity to a quantifiable model of some sort. 

Is Culinary Creativity Toast? 

You might be wondering what all that has to do with the signature dish you were visualising, before entering this conceptual woodwork. Well, while changing the conceptual paradigms that affect meta-level scientific or industrial discourse on food, AI has also started doing the sheer cooking.  

Carlo Cracco, in fact, isn’t alone in inviting AI into its kitchen. The Culinary Institute of Barcelona is already adapting its educational offer to AI’s presence: during the last academic year, students at the PCAC programme for high-cuisine chefs were required to turn into reality images of dishes that AI had created after receiving a textual description of them (via ChatGPT and MidJourned). In other words, while the initial prompt for this specific exercise was still human, top-class chefs of tomorrow in Barcelona are already learning how to execute what AI models tell them to cook. 

The same principle will be applied to operations of WOOHOO, a Dubai-based restaurant which will open in September and will be entirely run by a Large Language Model named Aiman—a cocky combination of the words AI and man. The AI Chef even gave an interview to Reuters a few weeks ago, explaining its gastronomical philosophy based on a combination of international flare and eagerness for recycling. 

In all this, the implication that, at some point, the creativity factor will have entirely migrated to AI agents does give chilling food for thought; the fact that a human hand might have turned the stove on does little to console.

Statement

AI has entered the food-making realm. Yet what may begin as a useful aide-mémoire for seasonal produce or a tool for avoiding waste could end up dictating taste itself—flattening centuries of geographical or regional diversity into algorithmic averages. If gastronomical skill is reduced to data points, and tradition to a quaint footnote, the chef’s craft risks becoming execution without imagination. The deeper question, then, is not whether AI can cook, but whether it should be allowed to decide what cooking means in the first place.