Academic integrity offices across the country are making disciplinary decisions based on AI detection scores. Students are appealing grades and fighting academic misconduct charges on the strength of what those scores say. And the tools producing those scores, like Turnitin, GPTZero and the others, are operating with a confidence that my testing didn’t find them to have earned.
I submitted both AI-generated text and human-written text to every major detector I could access, documented the results, and reached a conclusion that anyone using this technology to evaluate student work should probably read before they do it again.
Detectors…
Read the full article at CNET.COM
