Turnitin false positive rate is getting worse (August data)
So I’ve been testing Undetectable AI for the past few weeks and wanted to share my findings with everyone.
I know a lot of people have strong opinions on this but I think the data speaks for itself. I tried to be as objective as possible and test everything under the same conditions. Your mileage may vary of course but these numbers are hard to argue with.
The thing that surprised me most was how different the results were compared to what I expected. I went in thinking one thing and came out with a completely different perspective. I ran my tests across multiple samples to make sure I wasn’t getting fluky results and the pattern held up consistently.
What do you all think? Am I off base here or does this match your experience too?
5 Replies
Join the discussion.
Log In to Reply
Solid analysis. I've been testing Claude myself and got similar numbers. The consistency is getting better but still not where it needs to be.
The part about bias in detection really resonated with me. I think that's the most overlooked aspect of this whole debate.
Great point. I'd also add that the technology is changing so fast that policies can't keep up.