What Artificial Intelligence Changed About Being Recognised

For most of the history of photography, a face in an image was simply a face. It could be recognised by someone who already knew the person, and otherwise it was effectively anonymous. That is no longer true, and the change is more consequential than it first appears.

What changed is not that cameras improved. It is that a face has become something that can be searched. A single photograph can now be used to find other images of the same person across an enormous range of sources, with a confidence that was not possible even a few years ago. A face has become, in effect, a key.

This removes a protection that most people never knew they relied upon. A person could appear in the background of someone else's photograph, in a directory, in an old article, in an image from an event, and each appearance stood alone. They were not connected to one another, because connecting them required knowing in advance that they were the same person. That work is now automatic.

The consequence is that a person's images, taken together, describe far more than any one of them does. An image carrying a location, joined to an image carrying a name, joined to an image carrying a date, becomes a record of where someone was and when. No single image was sensitive. The ability to assemble them was what changed.

For someone whose circumstances make them of interest, this is significant. It is no longer enough to consider what a particular photograph shows. The question is what every image of a person, wherever it sits, becomes once it can all be drawn together by anyone with an ordinary tool and a single starting picture.

The first step, as always, is knowing what exists: which images of a person are out there, where they sit, and what they reveal when they are read together rather than apart.