Yesterday, I interviewed a tech guy who told me something too important for me to wait until the podcast comes out for you to hear it:

Many free AI detectors have especially high false positive rates because they make money selling software that "humanizes" writing. 1/2

They tell students papers sound like AI — even when they're the students' own writing — so the company can scare them and sell them a product.

This goes beyond the general problem that even good AI detectors aren't perfect and shouldn't be the sole reason to accuse someone of writing with AI.

0
0
0

If you have a fediverse account, you can quote this note from your own instance. Search https://zirk.us/users/grammargirl/statuses/115935154609937486 on your instance and quote it. (Note that quoting is not supported in Mastodon.)