Identity lock explained: faces, products and garments.

The short answer

Identity lock is the practice of describing a subject once, in writing, and reusing that description word for word in every generation so it stays identical. It applies to anything that must not drift: a face, a product, a garment. Our earlier note covers why digital humans need it; this one covers where else it matters and how we check it holds.

Why AI subjects drift

Image models treat each generation as a fresh draw. Unless the subject is pinned down, small details wander between images: a label moves, a seam changes, a skin tone shifts. Across a set, the subject quietly becomes several different ones.

Products

For AI photography we write a product identity profile covering shape, material, colour, label and the details that must not drift. The setting, light and props change while the product stays identical. Your logo and packaging are used exactly as you supply them.

Garments

For AI fashion imagery each garment gets a profile covering colour, cut, fabric, trim and any print or detail. The model is locked too, so one person wears the whole collection. The garment is never redesigned.

Compare against the reference

Every output is checked against the reference images and the written profile. When a result drifts, the reference is re-attached and the subject is treated as a locked asset instead of rewording the prompt.

Upscale without invention

Stills are upscaled without invented detail and graded once as a set, so the finishing steps do not introduce changes of their own.

Where it fits

Identity lock is the reason one digital human can appear in every ad, one product can look identical across a pack, and one collection can sit on one model. It is part of every SLIIQQUE service, and the part we are least willing to rush.