3 comments

[ 1.7 ms ] story [ 15.8 ms ] thread
I just feel the whole process of emerging AI and then AI detectors is dehumanizing. If someone were to use it on my work, I would feel another layer of the machine between me and that person, another layer of distrust. I think it's sickening that we've come to this stage in the world where people try and cheat (on university essays and projects and job applications) by wasting energy and then people have to use more energy to detect it. It feels futile.
Is the em dash in the title meant to be ironic?
I feel we should have benchmarks for each model to show if they are capable of being undetected as this is a real world use case (able to blend in genuinely with humans and even pretend to be them).