Why it matters

Trait prediction and identity matching answer different questions. Neither a predicted trait nor a generated composite is an exact photograph.

Watch the wording of the result

“Consistent with a predicted trait” conveys a different claim from “this is what the person looks like.” The first leaves room for uncertainty; the second suggests a degree of specificity that needs much stronger support. A visual composite should not erase that difference.

When comparing methods, ask whether they predict the same trait under comparable conditions. Avoid carrying a performance claim from one output to another. A reader should be able to identify both the useful information and the part of the picture that remains unknown.

Sources