Provost’s colloquium on AI raises questions over the future of academic integrity
JJ McKinney
February 21, 2023
Editor’s Note: Read the Spanish version of this article here.
Rick Miranda launched the first event in the Provost’s Ethics Colloquium series in 2016, and the events have run roughly once a semester ever since, except for a notable break during the pandemic.
Ad
The goal of the events is to promote and facilitate cross-disciplinary and cross-college discussions on the ethical dimension of issues that face CSU and academia generally.
The Feb. 16 event was focused on the academic impact of ChatGPT.
The goal was to discuss both the opportunities and challenges the new technology presents, bringing panelists from across the university to discuss it in depth.
The event started with opening remarks by CSU Interim Provost Janice Nerger, who, having left her prepared speaking notes at home, presented two sets of opening remarks both written by ChatGPT, one with a positive spin on the impact of ChatGPT on academics, the other with a negative spin.
The remarks were well-written enough that it would have been hard to determine whether or not they were written by the provost herself, as Nerger joked.
The discussion among the panelists began with a focus on the potential benefits ChatGPT presents to academia and in the classroom.
There was much discussion around the optimism panelists felt toward the development of ChatGPT, but few possible benefits were brought up without a qualifying statement recognizing the negative side effects of using chatbots.
One of those benefits was the potential outlined by Steven Lovaas, the chief information security officer for the CSU System. He said the chatbots could be used in mock cyber-battles to develop tools and strategies to even the playing field against technologically advanced adversaries using artificial intelligence in their cyber attacks on the university.
The conversation eventually pivoted to focus more directly on the concerns the new technology raises surrounding its impact on academia.
Ad
“So the 800-pound gorilla in the room with regard to challenges, I think, is cheating and plagiarism,” said Nikhil Krishnaswamy, assistant professor of computer science at CSU specializing in natural language processing.
The discussion of the panel covered a wide variety of subjects including the impact of ChatGPT on academic research, the equity concerns the adoption and use of the technology could pose in the classroom and how much effort and time it would take to try to detect work created by ChatGPT if it was even possible to accurately do so.
The audience asked the panel questions focused on how ChatGPT may perpetuate the profiling of students of color in academic settings and who actually owns the content generated by ChatGPT.
The overall theme of both the panelists’ discussion on ChatGPT and the questions raised by the audience, over half of whom were university instructors, was the concern of the detrimental potential of the technology.
Some of the concerns discussed prompted images of a bleak future.
“If you are imagining a classroom in which students are feeling overwhelmed, overworked, stressed by deadlines and use ChatGPT to submit work,” Lovaas said, “That (work) is presented to a faculty member who may be stressed by deadlines, overworked, using AI tools to do some of the basic administration of the curriculum in the classroom and then being asked to spend time using an AI tool to determine whether that submission was correct. You have a vicious cycle that could end with no one actually in the classroom. Bots teaching bots evaluated by bots.”
Paul DiRado, a senior instructor in the department of philosophy at CSU, discussed how he caught three students generating their work for his class using ChatGPT last fall by finding the papers that read like they had been written by a committee. This sparked a greater concern for him surrounding the flattening of language in a world with frequent use of AI chatbots like ChatGPT.
The panel also discussed the idea of a world with AI haves and have-nots and the the unfair distribution a powerful tool like AI chatbots would create for the university and in society.
Krishnaswamy said the defining line in his eyes between ChatGPT becoming a beneficial tool or becoming a harbinger of mediocrity is in the human-created parameters surrounding the specific uses of AI tools, something he said would need intensive discussion going forward.
After the panel ended, Matthew Hickey, associate dean for research and graduate studies in the College of Health and Human Sciences and a university distinguished teaching scholar, mentioned how pleased he was at the success of the event.
Nerger said she also was very pleased with how the event went, commenting on how important the Ethics Colloquiums are and how they allow academicians to do what they love: talk about important issues.
Nerger also said the event was just the beginning. CSU plans to have at least one more event this semester, likely on a similar topic but with greater student engagement.
“You know what’s missing is the students, and so the next one will have the student perspective too,” Nerger said. “We wanted to do this because we are all new to this, and we didn’t even know where we stood, and now we (have to) bring the student perspective in.”
Reach Grant Coursey at news@collegian.com or on Twitter @grantcoursey