False negative

A false negative is when a detector fails to identify AI-generated text and labels it as human. It's the less reputationally-costly error than a false positive, but it's the more common one — especially as humanizer tools (see humanizer) get better and as new model families ship faster than detectors are updated.

False negatives are why no production system should treat the absence of a detection signal as proof of human authorship. The detector telling you "this looks human" is not the same as "this is human." Combine detection with other signals (style consistency, edit history, identity verification, source attribution) when the stakes are high.