AI in Food and Cooking: How Technology Is Changing What We Eat

AI Has Entered the Kitchen

Food might seem like the last place AI would show up — it is personal, sensory, and deeply cultural. But AI is quietly changing almost every part of the food chain: from the farms where ingredients grow, to the apps that deliver it, to the kitchens where chefs experiment with new flavours. Here is how it works.

IBM's Chef Watson: AI as Recipe Inventor

In 2014, IBM Research created an AI called Chef Watson that could generate recipe ideas humans had never considered. Watson was trained on a database of 10,000 existing recipes and a chemistry database of thousands of flavour compounds — the molecules that give ingredients their taste and smell.

Watson's insight: ingredients that share flavour compounds often taste good together, even if no human chef had tried the combination. The AI generated recipes like Austrian chocolate burrito (chocolate and black beans share cocoa compounds) and Bengali butternut squash soup with cardamom (both contain similar terpene compounds). Not all combinations worked — but many surprised professional chefs with how well they did. (Source: IBM Research, Florian Pinel et al., "Optimizing the sensory properties and nutritional profile of dishes," 2015)

This approach — using chemistry to predict flavour compatibility — is now used by food companies including Firmenich and IFF (International Flavours and Fragrances) to design new products faster than traditional trial-and-error methods.

AI on the Farm: Predicting Yields and Reducing Waste

Before food reaches a kitchen, it has to be grown — and agriculture is one of the most data-rich industries in the world. Satellite imagery, soil sensors, weather stations, and crop monitoring drones generate millions of data points per day on a modern farm.

AI systems like those built by The Climate Corporation (owned by Bayer) and Indigo Agriculture analyse this data to predict crop yields weeks before harvest, identify areas of a field that need more or less water, and detect plant diseases from aerial photos before they spread. According to the Food and Agriculture Organisation of the UN, up to one-third of all food produced globally is lost or wasted. AI-assisted precision agriculture is one of the most promising tools for reducing that figure. (Source: FAO, 2019; The Climate Corporation)

Your Food App Knows You Better Than You Think

Food delivery apps — Uber Eats, Deliveroo, Swiggy, Zomato — use recommendation AI to predict what you want before you search for it. The algorithms consider your order history, time of day, current weather, what others in your area are ordering right now, and how long it has been since you last ordered from a specific restaurant.

This is collaborative filtering (the same technology behind Netflix recommendations) applied to food. The more you use the app, the more accurate the predictions become. Some apps now show you personalised "predicted rating" scores for restaurants you have never ordered from, based on the overlap between your past choices and other users who have tried those restaurants. (Source: Uber Eats engineering blog; Deliveroo Data Science team, 2021)

AI and the Future of Protein

One of the most ambitious food AI projects involves designing new proteins. Companies like Perfect Day and Remilk use AI to design the precise genetic sequences for microorganisms (usually yeast or fungi) that produce dairy proteins identical to cow's milk — without any cows. This is called precision fermentation, and it lets food scientists programme biology to produce specific ingredients.

Impossible Foods and Beyond Meat use AI to analyse thousands of plant protein formulations and identify which combinations best replicate the texture, colour change, and fat behaviour of real meat when cooked. The AI dramatically speeds up what would otherwise take years of manual food science experiments. (Source: Perfect Day, Impossible Foods R&D documentation)

Fighting Food Waste with AI Cameras

Winnow, a company based in London, installs AI-powered cameras and scales above restaurant kitchen bins. The system photographs and weighs every item of food thrown away, then uses computer vision to classify what is being discarded — chips, salad, bread, vegetables. Over time, it identifies patterns: which dishes generate the most waste, at which times of day, and whether prep quantities should change.

In trials with IKEA restaurants and hotel chains, Winnow's system reduced food waste by up to 50% — saving thousands of pounds per kitchen per year and significantly reducing carbon emissions. (Source: Winnow case studies, 2020-2022)

AI cannot eat food. But it turns out AI is very good at watching what we throw away — and helping us throw away less of it.