■ MODERATE RISK ■ Hospitality & Food
Not the chef, though a good chunk of the kitchen brigade below them is fair game. Cooking at volume is being automated at the assembly end while the creative and managerial end stays human.
“AI generates recipes. But tasting, plating, and the chaos of a kitchen rush? That's human artistry.”
Our AI replacement risk score — how we score jobs
Being a chef is two jobs wearing one jacket. There is the creative half — developing dishes, tasting and adjusting, sourcing produce, designing a menu that is both interesting and gross-margin viable — and the operational half, which is running a kitchen through service: calling the pass, sequencing tickets so a table's four mains land together, managing food cost and waste, hiring and retaining a brigade in an industry with brutal turnover. Neither half is a recipe-generation problem, which is the only part AI does convincingly today.
Automation is real but concentrated in high-volume, low-variance production. Robotic fry stations, pizza assembly lines, automated wok and bowl systems, and rice and noodle machines are deployed in fast food and fast casual, where the menu is fixed and the throughput justifies the capital. Behind the scenes, AI is doing genuinely useful work in demand forecasting, ordering, waste reduction, rota building, and menu-engineering analysis of what actually sells at what margin. Commissary production and central kitchens have been deskilling line work for decades; robots accelerate that.
What resists is the physical variability and the taste feedback loop. Ingredients differ batch to batch; a sauce needs adjusting because today's tomatoes are sharper, a fish needs thirty seconds less because it is thinner at the tail. Plating is manual dexterity plus aesthetic judgement. And restaurants are experiences, not calorie delivery — diners pay a premium precisely for the human authorship on the menu. Our risk score of 28 reflects an industry where the commodity end of cooking automates hard while the chef, especially in independent and premium venues, remains the reason the room is full.
Automatability: our editorial assessment of current and near-term AI capability
Transformed but not eliminated by 2040. Through the late 2020s, robotic production spreads across quick-service and chain casual dining, thinning junior kitchen roles and central prep. Back-office automation of ordering, forecasting, and scheduling arrives faster and touches every venue. Independent and high-end restaurants change least — their economics depend on differentiation, and a machine-produced menu is the opposite of differentiation. Expect fewer commis roles and a slower path up the brigade.
In fast food and fast casual, increasingly yes — fryers, grills, bowl assembly, and beverage stations are being automated where the menu is fixed and volumes are high. What is being replaced is the line cook executing a standardised item, not the chef designing a menu. Independent restaurants rarely have the volume or capital to justify the machinery.
It can produce plausible, competent recipes and is a useful ideation partner for combinations you had not considered. It cannot taste, so it cannot tell you the dish is a third too sweet or that the texture fails. Development is iterative tasting; a model that has never eaten anything gives you a starting draft, not a finished dish.
Moderately, with a caveat: the security is at the top and thin at the bottom. Head chef, sous, and specialist roles remain in demand, while entry-level prep and line positions face automation in chains and permanent wage pressure everywhere. The route in is narrowing, so get through the junior years with a plan and prioritise kitchens that actually teach.
Palate, speed, and consistency remain the fundamentals, but add commercial literacy and a specialty — butchery, fermentation, pastry, a regional cuisine — that gives you something a chain kitchen cannot buy off the shelf. Also learn to manage people; the shortage in professional kitchens is of competent leaders, not of bodies.