In conclusion—and I will delve into this later—your beloved AI detector may deliver a bunch of false positives, or Type I errors.
For example, if every word in a post is spelled correctly, that’s an obvious sign the text wasn’t written by a human—correct? In the ever-expanding world of virtual communication, correct spelling is a dead giveaway of non-human content—as is the use of characters unavailable on a standard keyboard. Motörhead made a bunch of £ and € despite not being real. As the band never said,
“Timothy Leary’s dead
No, no, no, no, he’s outside, looking in”
(I had to include one hallucination in this post.)
Use MFAID (multi factor AI detection) to increase accuracy when you claim to detect generative AI.
(Timothy Leary image public domain; lyrics from the Moody Blues, “Legend of a Mind”)
If you want to delve into so-called signs of generative AI writing, see
Bredemarket 