What We Actually Mean When We Talk About Taste in Design.

Everyone in design seems to agree at the moment that taste is the thing AI can't replace, and when we recently asked our teachers, students, graduates and friends across the industry which skills matter more now, almost all of them said taste and judgement.

Ryan McNally, one of our graduates who now runs his own studio, gave that answer and then immediately asked the question I suspect a lot of people are quietly asking themselves, which was whether taste could be trained into a machine, before answering it himself with a fairly resigned "probably". It's a fair challenge, and I think the only honest way to meet it is to be much clearer about what we mean by taste in the first place, because the word gets used so often and so comfortably that it has started to mean very little at all.

So I want to say what I mean by it, because I don't mean the thing you do with your tongue and I definitely don't mean "tasteful", that polite and slightly anxious idea of good taste in which everything ends up looking like everything else, all the same muted palettes and the same safe sans serifs and the same mood boards shared and reshared until nobody can remember where they started. Taste, for me, is the sum of everything you have ever experienced through every one of your senses across your whole life: the smell of the place you grew up, the songs that were playing when something mattered, the signage of a town you only visited once, the textures and conversations and disappointments and small accidental beauties that pile up inside you without your permission and come back out, often without you noticing, every time you have to choose one thing over another. What makes it worth anything is not whether it produces something pleasant but whether what it produces adds to visual culture in a way that carries real value rather than simply adding to the noise, and there is a great deal of noise around at the moment.

That distinction matters because "tasteful" in the narrow sense is exactly what AI is good at. A system trained on everything that already exists is very well equipped to produce the average of it, and the average is where polite good taste has always been heading anyway, which is why Sophie Duncan, one of our recent graduates, described AI as making "more of the same sameness at accelerating rate". On that level I think Ryan is partly right, because the surface of taste, a style or a palette or a layout that feels current, can be imitated and already is, and we wouldn't be doing anyone any favours by pretending otherwise.

What can't be imitated is the life behind the choice. One of our current students put it better than I could when she wrote that a human designer can turn something they have lived through into a design decision, while AI can only ever work with accounts of that experience, and that seems to me to be the whole thing. Taste in the sense I mean is particular and uneven and impossible to repeat, it comes from having been somewhere and noticed something that nobody has written down yet, and as Spencer Harrison, one of our teachers, said, AI is defined by the past while designers get to create the future, which only ever comes from people who are paying attention to the present.

The part that gives me most hope is that this kind of taste is not a gift handed to a lucky few, because if it is made of experience then you can go and get more of it, and you can get much better at noticing and using what you already have. That is most of what happens in a good studio and most of what we try to do in our classes, looking hard at a lot of work, making a great deal of it, having it critiqued by people who care and being asked again and again to explain why you made the choice you made, until the explaining becomes instinct and the instinct becomes yours. It is also why I would tell anyone starting out to get away from the screen as often as they can, to make things by hand, to go places and read things and listen to music they wouldn't normally choose, because all of it goes in and sooner or later all of it comes back out.

I don't think the question of whether taste can be trained has a neat answer, and the people we surveyed were honest about the real pressures on the industry right now, which we will come back to over the coming weeks.

But I am fairly sure that the designers who matter in five years' time will not be the ones who learnt to produce the most tasteful version of what already exists, they will be the ones whose work could only have come from them, which seems to me a pretty good reason to start, or to keep going.