Last year, students submitted more than 22 million papers that may have been written by AI tools, according to new data published by Turnitin, a software service that checks papers for plagiarism.
Turnitin says its plagiarism detection tools have uncovered millions of papers that may contain large amounts of AI-generated content. report by wired.
Last year, the company launched an AI writing detection tool and trained it on a large number of papers written by high school and college students as well as known AI-generated content. Turnitin says its detection tools currently have a false positive rate of less than 1% when examining entire documents.
Turnitin says its detection tools analyzed more than 200 million papers submitted by students, and found that 11 percent of them appeared to consist of at least 20 percent written language generated by AI, and 3 percent of papers appeared to be composed of at least 20 percent AI-generated written language. He said he found it to be at least 80 percent. Written by AI.
These kinds of tools, like OpenAI’s ChatGPT, have become a big issue in academia as students increasingly use them as their go-to source for cheating.
As Breitbart News reported, a study published last year found that 17% of Stanford University students admitted to using ChatGPT on their final exam. A student in a Florida high school’s elite education program has been accused of cheating by using ChatGPT to write an essay.
Educators want to hold students accountable for their misbehavior, but accurately detecting the use of AI in the creation of assignments can be difficult.
In particular, AI detection tools can pose a risk of false positives for English learners.
As Breitbart News reported, a Stanford University study last year found that AI detection tools were biased against non-native English speakers and falsely accused them of cheating.
Turnitin claims to have trained its plagiarism detector based on the writing styles of both English learners and native English speakers, and an October study found the software company was one of the most accurate AI language detectors. Wired reported that it is listed as one of the
Annie Chechitelli, Turnitin’s chief product officer, suggested using the software not as a way to pass or fail students, but as a way for teachers to start conversations with students about the nuances of using generative AI. .
“People don’t really know where the line should be,” she says.
you can Follow Alana Mastrangelo Facebook and on X/Twitter @ARmastrangeloand further Instagram.
