Guides
How AI content detection actually works, what the evidence can and cannot support, and where the common advice is out of date.
How to tell if an image is AI-generated
Check the file's embedded metadata first — C2PA Content Credentials, EXIF software tags, and the PNG text chunk that diffusion tools write the prompt into. Visual inspection is unreliable in 2026 and should be the last resort, not the first.
Updated September 2026Why AI text detectors are unreliable
Text detectors measure statistical regularity, not authorship. They cannot distinguish a person who writes plainly from a model, they misjudge non-native English writers more often, and light editing defeats them — so their output should never be used as evidence of misconduct.
Updated September 2026Does Turnitin detect ChatGPT?
Turnitin does have an AI writing indicator, but it estimates statistical regularity rather than proving authorship — and after it revised its false-positive figures upward, several universities including Vanderbilt disabled it rather than act on its scores.
Updated September 2026What are Content Credentials (C2PA)?
Content Credentials are a cryptographically signed manifest embedded in a media file recording what created it and what was done to it since — the strongest available evidence of origin, and also the easiest to destroy, because saving a screenshot removes it entirely.
Updated September 2026