My problem with things like GPT-Zero and its like is that if the writing has any modicum of decent use of language or grammar, it's AI. Because people are dumb. Ninety percent of college-level papers that can be proven to be human-written get flagged as AI-written.
You are viewing a single comment's thread from:
I have not tested my master's thesis in gpt-zero, but some of my other writing often gets "20% AI" scores when I check it.
I wonder if there are actual linguists working on the detection models. I doubt it, but I cbf doing that research, because my .. wonder gets me distracted from whatever it is I am doing at the time. :D
Look! Squirrel!