By SHANNON O. WELLS
The arrival and embrace of generative artificial intelligence tools at Pitt and other universities late last year was — to say the least — swift, pervasive and more than a little disruptive.
While the ChatGPT program gives students a fast, convenient way to synthesize complex information and create informed, if somewhat stilted-sounding, essays, firms like Turnitin marketed software (see related story) to help faculty detect misuse of generative AI in their assignments. This resulting tech-tool “arms race” put Pitt’s Center for Teaching and Learning on high alert.
“We did not have the luxury, unfortunately, of being able to sit back and think about it,” said John Radzilowicz, the Teaching Center’s interim director of teaching support. “We had to jump in headfirst.”
When image-based AI software emerged in spring 2022, the center’s staff had communicated with other institutions to help prepare Pitt faculty for generative AI-related problems. The emergence of ChatGPT into the academic landscape, however, was faster-spreading than many expected.
“We got this heads-up just a week or two ahead of time that (text-based) AI bots were coming out — ChatGPT, specifically — in November, so we began to scramble very quickly,” Radzilowicz admitted. “Even with that little bit of a heads-up, that wasn’t much. We didn’t think it would be that soon.”
Recognizing the need to provide useful information and guidance, Radzilowicz and his learning center colleagues started work on a web page of resources for faculty and staff. “We got out in front of it fast enough that in those first couple of months, if you did a Google search for generative AI and higher ed, in your top 10 Google hits you would have had the Teaching Center,” he said. “So we became a bit of a model for some other folks who are developing their things.”
Since that initial rush, which triggered a torrent of discussion and forums on what generative AI means for the future of teaching, learning and higher education, the University formed an Ad Hoc Committee on Generative AI in Research and Teaching.
Initiated by former Provost Ann Cudd, Senior Vice Chancellor for Research Rob Rutenbar and current interim Provost Joe McCarthy, the 29-member committee is co-chaired by Radzilowicz and Lisa Parker, professor of bioethics and director of Pitt’s Center for Bioethics & Health Law.
The committee’s focus includes:
-
Identifying topics where guidance is needed regarding generative AI applications in research, teaching and learning.
-
Conducting benchmarking with peer institutions regarding policies or guidance on Gen AI uses, particularly research uses to complement what the Teaching Center has already gathered on educational uses.
-
Mapping the landscape for uses of Gen AI at Pitt, including research and instructional uses.
-
Crowdsourcing and creating Gen AI resources, initiatives, activities — such as short courses — and programming.
- Identifying key areas in which Pitt can position itself as a leader in the Gen AI arena.
The committee’s primary goal, Radzilowicz explained, “is to create an umbrella under which all of these efforts can find a home. It’s not so much about dictating policy and approach. Though policy at some point might be made, that’s not going to be us. We may make some recommendations, but what we’re really trying to do is connect people, because there are already great things that are starting to happen.
“Every day we learn about more (projects), and we just want everybody — much the way large language models work —connected through nodes, so that the right hand knows what the left hand is doing,” he added. “So we can leverage all of that great work that’s already happening, leverage the resources we have (into) an approach that is very well informed about what’s happening not only at Pitt, but in other locations.”
Two-pronged committee
The committee evolved from conversations Parker and Rutenbar had following a Faculty Assembly presentation on Gen AI last year. Agreeing that the Gen AI explosion called for guidance pertaining to research and ethics, Rutenbar approached then-Provost Ann Cudd, who soon brought her successor, Joe McCarthy, into the discussion.
They decided to take a “two-pronged approach” encompassing research as well as education. “There was already so much going on in the Center for Teaching and Learning in terms of (supporting) faculty as they scramble,” Parker said, “not just to start the semester, but to take into account the existence of generative AI, ChatGPT and other similar technologies.
“The three of them met with us and kind of gave us the charge to assemble a committee,” Parker added. “They made a few recommendations (that snowballed) with some people recommending other people … to identify the topics where guidance is needed regarding generative AI applications in research AND education,” that includes the students’ learning process as well as teaching.
The approach to crowdsourcing and gathering resources from various University domains, including different campuses and units, is “much as the Center for Teaching and Learning has done for (non-AI related) educational resources,” Parker noted. “We hope to do that with regard to generative AI and resources more generally (through) initiatives, activities, programming, lectures and workshops and so on that are going on.”
Benchmarking with peer institutions regarding policies or guidance they have on Gen AI uses, particularly in the research context, will help identify key areas where Pitt can position itself as a leader in the field. “So, for example, the intersection with responsible data science, Michael Colaresi’s area and so forth,” Parker said, referring to Pitt political science professor and associate vice provost for data science.
Related to the Gen AI committee’s goals, the Research, Ethics and Society Initiative is pursuing a project to map Gen AI’s range of uses and approaches across various Pitt units. Project leader Ravit Dotan, a post-doctoral research associate with Pitt’s Center for Governance and Markets, will report the findings to the Gen AI committee, among others.
The committee plans to have an initial report prepared by December.
“You might think of (Doton’s) as a pilot project to map the landscape across eight different somewhat representative — certainly a range of units at Pitt — how they are using generative AI already, what their uses and their approaches to it,” Parker said. “So we hope, certainly by the end of the semester, to be making our early interim report back to Pitt Research and the provost.
“In a way, faculty shouldn’t be expecting too much from the committee right now,” Parker added. “We’re trying to identify kind of an agenda for guidance development … and amplify the efforts that others are already making.”
“Hopefully, they will see the effects of our work rather than seeing it directly from us,” Radzilowicz added. “And that will take time.”
Real-world tools
Reactions to generative AI from faculty at Pitt and elsewhere — at least initially — range wildly. There’s intrigue and even excitement about how to incorporate the bold technology into curriculum, as well as disdain and fear that generative AI applications give students a much-too-easy path to trade creativity and novel ideas for digitally sourced and packaged material.
Radzilowicz and Parker are among those finding a comfortable neutral ground from which to navigate the potentially revolutionary technology. Parker observed that where skills-based courses are concerned, “where you’re trying to teach the method of a discipline, you don’t want to send that over to, or allow generative AI to take over the doing of that, because students need to learn how to do it,” she said.
“Researchers need to learn how to write a background and significance section for their research proposal,” she added. “But maybe once you know how to do some of those basic things, then generative AI can be a useful tool to save some amount of time.”
Gen AI, Radzilowicz acknowledged, is simply the latest in a long continuum of perceived technological threats to academic purity.
“We’ve been here before, those of us of a certain age have,” he said. “We started from calculators to word processors to computers to the internet itself, and on and on. And this does not have to be an existential crisis at all. Instead, it’s a new tool, and what we need to do is grapple with it, learn to manage it. And like any good tool, use it to the benefit of not only the faculty, but the students. Because we have a responsibility to the students.
“If we turn around and just ban it and say you can’t use it,” he added, “well, what do (students) do when they graduate in a couple of years and they head out into the real world in the job market and suddenly discover that they have to use it as part of their work — and they don’t know anything about it?”
Shannon O. Wells is a writer for the University Times. Reach him at shannonw@pitt.edu.
Have a story idea or news to share? Share it with the University Times.