The Center for Computing & Data Science (CDS) adopted a new policy regulating the use of generative artificial intelligence in the classroom. CDS Associate Provost, Azer Bestavros, announced March 21 that the student-devised Generative AI Assistance Policy (GAIA) will take effect immediately.
The creation of ChatGPT, an AI chatbot capable of creating human-like dialogue, shocked the global computing community in November. Wesley Wildman, professor of philosophy, theology and ethics and computing & data sciences, was one member of CDS faculty who helped modify the school’s curricula in response to ChatGPT’s newfound fame.
Wildman introduced ChatGPT as a real-world computing ethics topic to his Data, Society & Ethics students (CDS DS 380), which covers real world case studies weekly that encourage discussion and debate.
During the first case study of the semester, Wildman’s students collaborated to come up with an AI-use policy that could be recommended to the CDS administration.
“[The students] said ‘don’t pretend that this stuff doesn’t exist,’” Wildman said. “[Students said] we really need to use this and don’t let us damage our skill sets … because we need this for our careers.”
Michael Lott, a sophomore in the College of Computing and Data Sciences and course assistant in Combinatoric Structures (CAS CS 131), said he’s seen students try to cheat on online exams by using AI tools.
“A lot of students … try to plug in problems into ChatGPT and get the solutions, and it’s not always right and it’s pretty frustrating,” Lott said
Nicole Liu, a freshman in CDS, said although ChatGPT can be viewed as cheating, the tool is useful and should not be banned.
“I do understand that professors want there to be a way for us to test that knowledge without losing touch with it, just to make sure we actually understand and grasp what we ‘copied’ [from ChatGPT],” Liu said. “I think that as long as you’re … not just running [problems] through it and then copying and pasting the answers directly, it’s fine.”
However, CDS faculty and administrators largely disagree with the popular rhetoric that using AI constitutes cheating.
The GAIA policy outlines that students may use AI tools, when professors allow, if they properly credit the tool and include an appendix showing where and how AI tools were used. The policy suggests that students will be equally rewarded for using AI tools “wisely and intelligently” as they would be for completing their work without AI.
Peter Golbus, a lecturer of computer science in CDS, said that the policy should be implemented differently depending on the class or major of the student.
Golbus said students in advanced software-engineering classes who already have foundational skills as well as non-CDS majors should be allowed to use ChatGPT.
“Whereas for the courses where the goal is to develop foundational skills, you may not use ChatGPT, because that would defeat the purpose.”
Golbus added that the CDS policy should emphasize that students should be cognizant of the goals of the assignment, and use AI “in a way that furthers rather than hinders those goals.”
The CDS policies will need to remain fluid, requiring regular updates in consideration of the fast speed at which AI technology evolves, Wildman explained.
One example of these changing policies is the requirement that students employ originality checks and AI detection software to their work, Wildman said.
“It’s very likely that [AI detection tools] aren’t going to work for much longer because generative AIs are getting so good at producing human-like text that there won’t be a basis to be able to detect it,” Wildman said. “Hopefully by then we’ll have gotten better at crafting assignments to evaluate knowledge that mean that we’re not as dependent on whether or not students choose to use generative AI.”
