Chat GPT and similar artificial intelligence chatbots, commonly referred to as A.I., have been growing in popularity across the nation. Professors, teachers, administrators and corporations have been facing the question of if this tool should be used. Some corporations have decided to embrace A.I. as a time-saving tool, while others have decided to completely ban its use. Despite the strong stance against A.I. taken by most schools and colleges, some students continue to use these chatbots to supplement their own learning or complete their assignments.
I bet it’s hard to believe that paragraph was written by A.I. — and that’s because it wasn’t. I spent an hour trying to make it look and sound like A.I.
While higher education tends to argue against the use of A.I., there is some pushback on completely banning the use of A.I. chatbots.
But this brings up an important question, something that is a hot topic among professors and teachers: how can you reliably determine if something has been written by AI? Even Turnitin, one of the most commonly used plagiarism detector softwares, is struggling.
Chief Product Officer of Turnitin Annie Chechitelli released a statement in May of this year in which she addressed problems with false positives and errors in interpreting the results. Turnitin claims to have a false positive rate of about 4%, but suggests that scores under 20% are too unreliable for a false positive to be ruled out when assessing for A.I. generated text.
In an interview with the Washington Post, Soheil Feizi, a professor of computer science at University of Maryland, stated these programs “have a very high false positive rate and can pretty easily be evaded.” Feizi added that these programs may show more false positives on papers written by people who speak English as a second language.
While higher education tends to argue against the use of A.I., there is some pushback on completely banning the use of A.I. chatbots. Proponents claim A.I. can be used as a tool to aid teachers, professors and students. A review article from 2020 suggests A.I. could be used to grade tests and has been used to create personalized study material for students.
A nationwide self-report survey by Bestcolleges found that 22% of college students admit to having used A.I. to complete assignments and 32% of students plan to continue using A.I. to complete assignments.
This rise in use of A.I. has raised questions about what can be done to prevent students from passing off the work of these chatbots as their own. Some suggest that pushing A.I. completely out of education might not be the answer. It might be possible for A.I. to be used to supplement writing and reading comprehension, similar to the way calculators are used in the instruction of mathematics.
This rise in A.I. doesn’t necessarily need to be something scary. Embracing change by finding ways to incorporate new technology in ethical ways may be more beneficial to students and their careers. Students need guidance to learn how to use these tools properly so they can grow and bring new innovations into the world.
jgray11@ramapo.edu
Featured photo courtesy of Hatice Baran, Pexels
