Sometimes the best ideas for fashion don’t come from fashion.

Take Sarah Ezekiel’s story. After losing her voice to motor neurone disease, AI has just given it back - from only eight seconds of scratchy VHS tape. A Londoner who hadn’t spoken with her own voice for 25 years can now sound like herself again. Not a robot. Herself.

It’s moving. It’s human. And it’s a powerful lesson:

1. Technology should restore identity, not flatten it.
Fashion, like voice, is about self-expression. If AI in fashion becomes generic — the same styling, the same copy, the same outputs — it strips people of individuality. The opportunity is to build tools that amplify uniqueness: your brand’s voice, your customer’s identity.

2. Tiny data. Big impact.
Eight seconds. That’s all it took to unlock something life-changing. In fashion, we obsess over “not enough data.” But fragments can be enough — if the intent is sharp. A few signals on preference, fit, or behaviour can deliver personalisation that feels magical.

3. Tech should deepen the bond
What made this story matter wasn’t the software. It was Sarah’s kids hearing their mum’s voice. Fashion should remember: the endgame isn’t slicker ops, it’s connection. Confidence. Belonging. Tech should deepen the bond, not dilute it.

4. Local voices, local styles.
Accents matter. Origins matter. They’re markers of who you are. Fashion faces the same risk as AI: homogenisation. Don’t let it happen. Preserve dialects of style, not just globalised sameness.

If AI can give someone their voice back after 25 years, it can help fashion find its voice too. Your choice: use AI to amplify individuality - or let it flatten identity into sameness.