Plagiarism and AI Detectors
By: Janet L Donavan
础产蝉迟谤补肠迟:听
I confess to adopting plagiarism detection without much thought. Our LMS has an integrated plagiarism detector and I 鈥渃hecked the box鈥 for plagiarism detection for years. My thought would help students to see their 鈥減lagiarism score鈥 in advance of submitting the paper and make adjustments. I design assignments using best practices to make them difficult to plagiarize (WPA 2017) and I have only identified a few cases of plagiarism in my classes using this software across thousands of students. In most cases, these were cases of sloppy attribution, or an overly aggressive algorithm citing common phrases as plagiarism. The main value is to encourage students to catch plagiarism before submitting work. This year, an instructor reached out for advice after being contacted by the plagiarism detection software provider asking for a student assignment to be sent to an instructor at another institution for review. Because the student saved the assignment with their name as the file name (which students are often instructed to do, for example 101_Donavan_Midterm1), the student鈥檚 name had been retained in the plagiarism system and revealed to this other instructor. I became aware of how student papers as well as instructor feedback are being saved in these systems and fed through plagiarism algorithms in ways that are sloppy at best at providing student privacy, with student or faculty names being shared by the system when they are in the file name.听
听