The next step for AI profiteers is creating problems for their jank AI to solve.
-
The next step for AI profiteers is creating problems for their jank AI to solve.
-
-
Howard Cohen for Harris/Walzreplied to Tzip last edited by
@Tzipporah @noondlyt [insert image of spider man pointing to second spider man who is pointing back]
-
MissConstruereplied to Howard Cohen for Harris/Walz last edited by
@hoco @Tzipporah @noondlyt
It’s already a thing. Schools and universities are buying AI bots to check students work. Of course they are wrong a lot, and students are being punished when they did nothing wrong, but hey, as long as a techbro gets his lambo, who cares about the destruction of knowledge and learning and the deprecation of the teaching profession. -
@MissConstrue @hoco @noondlyt yep. And of course they are more likely to flag work written by autistic students.
-
And English as second language students. And low income students. And anyone who writes sentence structures that are long, complicated, and have subclauses. Many doctoral thesises, written well before AI existed, will scan as AI written because they use complex language structure. Using AI as a gatekeeper is a travesty and should be stopped.
More Teachers Are Using AI-Detection Tools. Here's Why That Might Be a Problem
Students are increasingly getting disciplined for using generative AI, a new survey finds.
Education Week (www.edweek.org)
-
@MissConstrue @Tzipporah @hoco @noondlyt
These replies are kind of dancing around it, but as a computer scientist, I just want to spell it out clearly: supposed AI detectors are snake oil. Pure snake oil. Ethics of gatekeeping based on AI aside, the detectors simply do not work; the false positive rate is so high that they are worse than useless. Unsafe at any speed. Do not use.
