The photo can be real. The story attached to it can be wrong.

Split the story into testable parts

Make a short list: who is shown, where, when, what is happening and who published the claim. Leave a blank beside each item until you have a source. This prevents a plausible location from accidentally becoming proof of the date or allegation.

Look for the exact image in earlier reporting and read the original caption. If the first useful match is an archive page, check whether its displayed date describes the event, upload or later update.

Work through a hypothetical mismatch

Suppose a crowd photograph is shared as this morning’s event. You locate the same frame in an older photographer’s portfolio. The useful finding is that the current date claim conflicts with that earlier publication. You do not need to establish that the image was AI-generated to explain the problem.

1. Save the current caption and the page where it appears.
2. Locate an earlier use with a readable caption and date.
3. Compare the same frame, not merely a similar-looking crowd.
4. State exactly which claim the earlier record contradicts.

Keep file evidence in its proper role

Inspecting the original may add saved timestamps, editing records or production clues. Record their field names and avoid treating editable metadata as an independent eyewitness.

For readers, show the two relevant contexts side by side in words: “The current post says X; the earlier credited publication says Y.” Link both when appropriate. This gives a concrete correction without overstating what the image tool established.