In London there was Ziad Ghanem, who made theatrical couture and cast his shows with the people the industry usually erased: older bodies, disabled bodies, every kind of face. I loved him for refusing to pretend fashion was only decoration, or only for one kind of person. He did it years before anyone filed it under "inclusivity." He did it because it was obviously right.
Here is the strange place that lands in 2026. Representation became brand risk and policy, and then AI added a twist I keep thinking about. Brands now generate "diverse" models with software, a synthetic face of any age, size, or background, made to order. So the inclusivity Ziad practiced with real, consenting human beings becomes a consent question with no human in it. When the diverse model is generated, who agreed to exist, and whose features were borrowed to build her? That is the exact gap likeness and digital-replica law is racing to fill, and that the EU AI Act’s transparency rules begin to touch by requiring that synthetic media be disclosed.
Ziad was defending the human in the picture. That defense is becoming the law, and the thing it now defends against is a machine that can fake the picture entirely. If you use AI imagery in a campaign, the questions to ask are the old ones he answered with real people: who is this, and did they say yes.
