AI Knows What You'll Buy: Fashion Lost Its Nerve
There is a particular kind of anxiety that spreads through an industry when it realises the disruptor it invited in for drinks has quietly moved into the spare room.
AI Knows What You'll Buy: Fashion Lost Its Nerve
There is a particular kind of anxiety that spreads through an industry when it realises the disruptor it invited in for drinks has quietly moved into the spare room. That is where fashion is right now with artificial intelligence — and the Vogue Business AI Tracker, updated with the obsessive regularity of someone who cannot decide whether they are excited or afraid, is the most honest document in the room precisely because it cannot quite decide either.
The tracker reads like a dispatch from a negotiation nobody officially called. On one side: the houses, the editors, the stylists, the people who have spent careers insisting that taste cannot be systematised. On the other: the tools that are getting very good at proving them wrong. Not because AI has taste — it doesn't, not really, not yet — but because it has learned to simulate the *output* of taste convincingly enough that the difference stops mattering to the algorithm deciding what you see next.
This is what gets lost in the conversation about AI and fashion. Everyone talks about it as a production question — faster trend forecasting, smarter inventory, AI-generated campaigns — when the real question is a cultural one. Who decides what is beautiful, and what happens to fashion's entire identity project when that decision gets handed to a system trained on what people already clicked on? Fashion has always told you what to want before you knew you wanted it. That was the power. The moment you optimise for revealed preference, you are not leading taste anymore. You are following it at scale and calling it innovation.
Donatella Versace built a house on the idea that fashion should feel like a provocation before it feels like an invitation. Dsquared2 never asked permission from a single focus group. The entire avant-garde tradition of this industry — the Margielas, the Kawakubos, the moments that make fashion worth taking seriously — was built on someone deciding that what people thought they wanted was far less interesting than what they hadn't imagined yet. That instinct is not something you can train a model on, because it lives precisely in the gap between data and defiance.
The tracker will keep updating. The tools will keep improving. And somewhere in a studio, someone will make something genuinely strange and genuinely new, and nobody's algorithm will have seen it coming.
That's the part fashion needs to protect. Not the aesthetic. The nerve.
