Can a machine understand the flavour of a moment?
Imagine a chef travelling to work on an ordinary Tuesday morning. On the train, she notices what people are wearing, overhears fragments of conversation, senses the mood of the city. Perhaps the sky is grey, the streets smell of rain, and everyone seems unusually quiet. None of this has anything to do with recipes. Yet by the time she reaches the kitchen, something has changed. The food she wants to create today may not be quite the food she imagined yesterday.
When I spoke with French chef Alexis Gauthier, he described something remarkably similar. A chef, he suggested, brings much more into the kitchen than knowledge of ingredients and techniques. There is also what he calls the flavour of the moment: the colours people wear, their conversations, even the perfumes encountered on the way to work. These impressions become part of the creative process. And when Gauthier began working with ChatGPT, he wanted it to consider precisely this dimension of cooking. Not simply which ingredients belong together, but what a particular menu should feel like, here and now.
But how could a machine possibly know that? AI can analyse fashion trends, weather reports, cultural conversations and thousands of restaurant menus. It can recognise patterns and suggest combinations that reflect a particular season or mood. What it cannot do is sit on that train, feel the cold through its coat or arrive at work carrying the residue of an unexpected encounter. And there is another complication: the same rainy morning might inspire one chef to create something warming and comforting, while another might long for something bright, sharp and defiantly summery. The moment has no single flavour.
Perhaps, though, understanding the flavour of a moment does not require experiencing it in exactly the same way. A good creative partner need not share our memories or sensations to help us discover something unexpected within them. Gauthier found that ChatGPT became more useful as he taught it to think more like him. But who is learning from whom in such an exchange? Is the machine discovering the mood of the world, or helping the chef recognise more clearly what he already feels? And could that difference be precisely where human–AI creativity begins?