A picture editor's job was never choosing something attractive. It was knowing that the mountain in the frame is the wrong mountain, that the uniform belongs to a different country's police, that the trees are in leaf and the story is set in February. It was a specialism built out of years of looking, and it was largely invisible, because its output is the absence of an error.
That made it an easy cut. A role whose value shows up as things not happening is a role that survives exactly as long as somebody senior remembers why it exists.
What replaced it cannot see
The replacement is a search box. A writer or a producer types a description into a stock library or an internal archive and picks from a grid of results, under deadline, with no particular training in what to check.
Stock search matches words to words. A photograph is indexed by whatever its uploader typed, and roughly half carry no description at all — so a query for a government office returns any building anywhere, and a query for a police car returns any police car in any country. The system is not wrong. It is answering the question it was asked, which is a question about keywords rather than about what the picture shows.
The failures are consistent enough to catalogue: right subject, wrong country. Right season, wrong hemisphere. A file image of an event that resembles the one being reported, presented without saying so. Each is invisible to anyone who does not already know the place, and obvious to everyone who does — which, on the internet, is a substantial number of readers.
Small newsrooms carry the sharpest version of this. The one-person video newsroom is reporter, camera operator and picture desk simultaneously, and the last of those is the one nobody trained for. The independent outlets that have done best have generally rebuilt some of the function informally — a second person who looks at every image before it publishes, which is not a photo editor and catches most of what one would.
There is a straightforward remedy that costs nothing and is unevenly applied: say when a picture is a file image. A landscape captioned honestly as a file photograph from a different region is a defensible editorial choice. The same photograph presented as though it shows the event is not, and the distance between those two is one line of caption.
The economics that removed the desk have not reversed and will not. What has changed is that the error is now cheap to catch and expensive to leave — a misidentified photograph gets found within the hour by someone who lives there, and the correction costs more credibility than the picture ever earned.
Topics media
