By SHANNON O. WELLS
In the current blink-and-you’ll-miss-it trajectory of artificial intelligence-related developments, discussion about the potential benefits and drawbacks of AI’s role and applications in society often boil down to an overly simplistic “good or bad” paradigm.
Michael Colaresi, the newly named director of the new academic Hub for AI and Data Science Leadership (HAIL) and strategic advisor to the provost, prefers a more nuanced view of the game-changing digital technology.
“I absolutely think there’s a middle ground. And I almost think, like with all things digital, it’s just so easy to give in to the illusion that there’s sort of one right answer because it’s abstract. And we can talk about things in abstraction,” he said. “But the reason there’s absolutely a middle ground is because the way we use tools across different tasks that … influence different decisions — that are going to be seen by and impact on different communities — all of those can have positive or negative values, depending on the context.”
Colaresi, who will transition from his current associate vice provost for data science role by January, said with AI, there’s “no one-size-fits-all, even if we’re talking about a gigantic, 100 trillion-parameter model. The power of these tools, of these different ways of arraying information and language when we’re talking about generative AI, is in their uses.
“It’s in what jumps out at people in both good and bad ways.”
Culling the good ways from the bad is among Colaresi’s new responsibilities as director of HAIL.
Intended to build on the foundation created through the Responsible Data Science (RDS) initiative and other fledgling AI efforts across the Pitt community, HAIL seeks to serve as an academic convening point for AI and data science efforts and “facilitate collaboration across teaching, learning and other academic operations,” a statement from Provost Joe McCarthy said.
“In this expanded leadership role, (Colaresi) will not only guide and coordinate University-wide work advancing Pitt as a leader in academic usage and development of AI, but he will also engage with other leaders across the institution — in research, operations, the health sciences and infrastructure — to promote a unified strategic vision that ensures our whole campus community is prepared to responsibly and ethically lead today and into the future.”
In an effort to realize a University-wide vision for “responsible and innovative AI applications,” HAIL represents what McCarthy called a “natural progression of the early successes of RDS” while leveraging the accomplishments of other “exciting, currently disparate,” AI-related efforts.
To fulfill a growing need for coordination of “complementary expertise and experience across campus” while sharing evolving best practices, the HAIL hub intends to:
-
Unify and strengthen existing AI- and data science-related activities across Pitt’s academic mission.
-
Convene and connect researchers, practitioners and academic units to support responsible, innovative applications of AI.
-
Engage with shared governance, especially with the Senate Educational Policies Committee, the Student Admission, Aid and Affairs Committee, and the Student Government Board.
-
Expand upon current RDS work, incorporating its people, resources and commitments while broadening the scope to reflect the “rapidly evolving landscape of AI” and data science.
Under Colaresi’s leadership, McCarthy said, “this new hub will advance Pitt’s academic mission, support our students, network together faculty and staff across disciplines, and contribute meaningfully to the broader societal conversation on the responsible use and development of AI.”
Responsible use
Colaresi, who was among “many different stakeholders” involved in developing HAIL, said he feels “very lucky to have been included in the discussions of what this needs to be.”
“It became clear that we were doing more and more AI-related things, because that’s what faculty, students and staff needed. I absolutely felt very listened to when speaking about what we could accomplish with HAIL, and in particular, moving discussions and dialog forward in the academic mission of the University,” he added.
“That was really, really important, (along with) making sure faculty and student voices and shared governance were connected to what was going on.”
Responsible use of data is a significant piece of the AI tapestry that connects the initial development of Responsibility Data Science@Pitt into HAIL, Colaresi emphasized, describing his new role as “continuing the momentum” of the RDS initiative.
“(This) wasn’t about data science for data science’s sake, or technology for technology’s sake,” he said. “It was always about moving decisions and values forward on the ground, and that’s something (we’ll bring to HAIL), too.”
It’s hard to think about AI and its various forms — including generative, AI predictive and AI agentic — without what Colaresi called “data science under the hood.”
“We’re really excited that Pitt sees data science as being part of the power of responsibly applied AI, and that we continue that trajectory.”
By intentionally linking, rather than replacing, what McCarthy called Pitt’s “diverse and highly impactful AI-related initiatives” spanning learning sciences, the humanities and practical application in areas including business, nursing, engineering, etc., HAIL ideally will catalyze a “research-practitioner community of practice” that accelerates innovation and promotes “responsible, human-centered AI.”
‘Terrific contributions’
While AI efforts at Pitt initially grew within domain-specific contexts, the next stage of development, McCarthy noted, requires a “coordinated network that allows expertise in one area to strengthen the whole.”
Elaborating on this, Colaresi noted that data science throughout the past decade, “was in so many places,” that it was “hard to be in an industry that wasn’t touched by data and modeling. And there was a certain speed and breadth that came with that.
“What we’re seeing with different types of AI tools is that speed, a new level of acceleration — that breadth, a new level of expansion — in how students and faculty and staff and the workforce are all engaging with different tools.
As a result, Pitt needs to be there “at the forefront of how to responsibly teach this, how to responsibly operate,” he added. “So it’s really exciting to be in this position to help coordinate all of Pitt’s already, I think, terrific contributions in this area.”
Although Colaresi said HAIL is unique to Pitt, he and his AI-focused colleagues have “definitely learned from other initiatives out there.”
“It’s hard to miss that other organizations have connected data science and AI in the academic space,” he said, noting the “very high-profile” Academic Data Science Alliance (ADSA) is now the Alliance for Academic Data Science and AI.
“A number of places have realized that the evolution of data science connects deeply with responsible AI development and uses. We have learned from that, and that’s highlighted that connection.”
There’s also “things to learn” from peer institutions that demonstrate strong leadership in data science and AI, including what Colaresi called a “terrific organization” at Ohio State University called the Translational Data Analytics Institute.
“I think (that’s) a wonderful model,” he said, also citing the University of North Carolina’s School of Data and Society and the University of Virginia’s School of Data Science.
“I mean, there are many models out there, but I would say HAIL is uniquely a Pitt institution.”
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.