Get the latest Syracuse news delivered right to your inbox.
Subscribe to our newsletter here.

Syracuse University’s Maxwell School of Citizenship and Public Affairs hosted eight panelists who shared how their respective fields of academia influence their perspectives on ChatGPT on Monday.

The panelists discussed the incorporation of the November 2022 chatbot — which is trained to produce written responses to user-submitted prompts – in the classroom. The event also included a demonstration of the AI’s tendencies and interactions with respect to different types of prompts.

ChatGPT has ignited controversy in academia and education over plagiarism concerns. At SU, administrators created and disseminated a resource document providing guidelines and answering frequent questions for instructors at the start of the spring semester.

Margaret L. Usdansky, the director of SU’s Center for Learning and Student Success, said during the panel that ChatGPT could help educators and students.

“I’m hopeful that artificial intelligence will really push us to think more carefully about what we’re teaching and why we’re teaching it and the value of it in a way that is helpful and transparent to students and motivating in terms of learning,” Usdansky said.

Usdansky said artificial intelligence programs like ChatGPT will ultimately make professors think more carefully about what they’re teaching and why, and acknowledged concerns over academic integrity.

Verena Erlenbusch-Anderson – an associate professor of philosophy in the College of Arts and Sciences – agreed with Usdansky, and emphasized that ChatGPT provides an opportunity for professors to ensure assignments are intellectually challenging enough that ChatGPT can’t complete them in place of a human.

Some panelists also expressed concerns about the software’s potential bias. Delali Kumavie, an assistant professor in the Department of Women’s and Gender studies who specializes in studies in science and technology, explained that she was impressed with the software when examining it for racial and gender inequalities.

“As someone who’s looking at Blackness and technology, when I began using Chat GPT I was curious about… the ways in which data virus replicates forms of racial gender inequalities but what surprised me was that it didn’t do that,” Kumavie said.

Bei Yu, a professor and Ph.D. program director whose research focuses on machine learning and natural language processing to improve information quality, highlighted additional concerns about ChatGPT’s potential contributions to misinformation.

“I noticed some troubling issues and concerns that quite often are unpredictable,” Yu said. “Even as inventors, the creators cannot fully control what it’s going to output.”

Panelists also discussed the economic pressures that arise with ChatGPT. Baobao Zhang, an assistant professor of political science, spoke about the pressure to use the software in order to remain employed and keep up with industries.

“What if your colleague who uses it produces three more papers a year and you don’t?” Zhang said. “Then you don’t get promoted.”

Zhang also addressed concerns surrounding full computerization of jobs on top of altered competition among humans, saying that even if AI is good at performing certain tasks, there may be certain circumstances where they want human judgment instead. Still, many jobs do face automation. In a March 2023 study, Goldman Sachs researchers found that generative AI could automate the equivalent of 300 million full-time jobs.

Ferdinando Fioretto, an assistant professor with a research focus in ​​AI, privacy and machine learning, discussed security and privacy aspects of the software, drawing attention to the importance and implications of training data.

On Friday, ChatGPT was temporarily banned in Italy after privacy concerns about its creator OpenAI unlawfully collecting personal data from users and its lack of an age-verification system for minors.

The software has also come under criticism from Twitter and Tesla CEO Elon Musk as well as from Apple co-founder Steve Wozniak, who were among more than 1,000 executives and AI researchers who signed a letter on March 22 calling to “immediately pause” production of increasingly dangerous AI models. On March 13, Open AI launched GPT -4, which is the most recent update of the ChatGPT text generator engine.

Looking ahead at the future of ChatGPT and other AI generators in academia, Erlenbusch-Anderson said that she hopes to remind students of the value of completing their own work instead of relying on software.

“In my own teaching, I’ll continue what I have been doing so far… such that it’s clear for students that if they cheat, they’re not harming me,” Erlenbusch-Anderson said. “They’re robbing themselves of an opportunity to actually learn something.”

Published on April 4, 2023 at 2:08 am

Contact Claire: cksamsta@syr.edu