By SHANNON O. WELLS
An initiative to examine and plan Pitt’s technology-based needs and anticipate marketplace changes in the years ahead is moving forward into a public-comment and final-draft period.
Mark Henderson, chief information officer, provided an update on the initiative to the Computing and Information Committee meeting on Jan. 27. The meeting also included an overview of the Computational Pathology AI Center of Excellence and a draft document with guidelines on using artificial intelligence (AI) platforms with various classifications of University data.
Henderson shared the latest on the technology forecast initiative.
“As you all probably remember from last year, we kicked off an initiative around a planning process that looked at University needs and forecasted changes in the marketplace over a five-year horizon,” he said, noting that Deloitte business services was brought on board to facilitate feedback-generating interviews. “We’ve gotten down to our final draft, which we have been socializing. It’s still going forward.”
Deloitte-generated infrastructure has been set up to facilitate public comment based on an initiative document endorsed by the Information Technology Action Committee and reviewed by most senior leaders on the chancellor’s cabinet along with “a couple of deans.”
“Work continues on that front,” Henderson said. “It’s pretty exciting.”
Some of the questions put forward include:
-
How do we think about the University five years from now? Ten years from now?
-
What are some of the emerging technologies and trends that will have some impact on Pitt?
-
How do we think about the future of health sciences and medicine?
-
What are the kinds of technologies and emerging technologies and systems and approaches that need to be undergirded?
“Being that we’re a university that the chancellor sees — particularly through the Plan for Pitt — as being innovative and forward leaning and a leader, this 10-year planning initiative speaks to and aligns with the Plan for Pitt 2.0,” he said. “So we have attempted to address some of those in this.”
Other questions the initiative addresses include:
-
How do we think about students and what students may be after?
-
How do we think about the marketplace and how we prepare students who come out of the University?
-
How do we think about teaching and learning?
-
What are the technologies and potential approaches that are emerging, and what are the requirements to undergird that and provide scaffolding to achieve it?
“These are things that we’ve touched upon there,” Henderson said. “There’ll always be a need for basic blocking and tackling. No one wants to hear a lot about innovation if you can’t keep the lights on and the trains running on time, so that’s recognized in this as well.”
With support from Deloitte, the initiative is moving into a public-comment period. The Office of the Provost’s website will feature a QR code participants can scan to take a survey and provide feedback.
As part of his CIO Report to the committee, Henderson also provided an update on the Pitt Information Technology’s 10-Year Horizon – Preparing to Meet Pitt’s Evolving Digital Needs initiative.
“We had some original conversations a couple of years back when (Amazon Web Services) came up with the notion of their CIC (Cloud Innovation Center). They didn’t see, and we didn’t see, kind of a connection in our earlier conversations,” he said. “And it’s probably just as well, because they wound up rethinking their notion of a CIC.”
AWS returned with a reinvention, with Pitt IT being one of their first in what Henderson called AWS’s “new thinking.”
“We’ve entered into a one-year agreement, which provides about $200K in funding, where we’ll be hiring students (who) will be trained in the use of AWS technologies and systems,” Henderson said. “And we’re coming up with a list of interesting projects for … the University across a number of areas, and aligning those initiatives with things that AWS believes to be emerging for them.”
Keeping PACE
In other news, Pitt IT has been working with the Department of Pathology, Dell and UPMC on supporting the Computational Pathology AI Center of Excellence (CPACE).
“Our partnership with Dell intersects with not only their technology and how we use it here at the University, but also they aspire to have a positive effect on a billion souls by 2030,” Henderson said. “Being that we do a lot of research and very interesting work in the health sciences and medical field, there’s overlap.
“Dell is interested in investing in those things that will have a tremendous impact on humankind per their ‘moonshot’ activity.”
Calling the center Pitt’s next iteration of its partnership with Dell, Henderson said the University has received a “fairly generous grant” from Dell about two years ago that supported some initiatives through the Center for Research Computing and the Pittsburgh Supercomputing Center, “and that’s been good.”
Recognizing Pitt’s progress, he noted, led Dell into a “second round.”
“We’re trying to put the finishing touches on the proposal,” which will include UPMC as part of the grant. “That’s because there’s interest in having some clinical information that will help feed some of the large-language models that will come out of CPACE. We’ve been working closely with our counterparts across the spectrum at UPMC for this.”
Henderson said the goal is to finalize the proposal by early February and announce something later in the month. “Interesting opportunities,” he said. “So hang on to your hats. More to come.”
AI and Pitt data
Brian Hart, policies and governance coordinator for Pitt IT, provided an update on a draft operating standards document prepared to assist those using University-based data with AI technology.
The document, which applies to Pitt users providing data to GenAI models and systems in connection with “University teaching, learning, research, business and other activities,” explains the need to limit use of “private and restricted University data,” or to not use private and restricted University data with public generative AI tools.
“This isn’t an AI policy in its largest sense,” Hart explained. “It’s not intended to cover any aspect of anything dealing (directly) with AI, which would be kind of impossible at this juncture, given the developments with AI and things changing so rapidly in that space.”
The operating standard describes and limits the types of generative AI models and software that can be used with various classifications of Pitt data. This standard also “provides user responsibilities for the use of Generative Artificial Intelligence,” which includes: 1) the deep learning models that produce content, such as text, images or video from patterns in data that are provided to the model; and 2) any systems that use GenAI models, like Chat GPT, Grammarly, Notion AI, Microsoft Copilot.
“Public GenAI tools” are considered those readily available on the general market to anyone, whether free of charge or on a paid-subscription basis. “University GenAI tools” are those “specifically approved for use with University data. Tools approved for use with University Private or Restricted Data interact with private University Large Language Models (LLMs) that do not share this data with other entities or the general public.
Pitt users, the document says, “are responsible to guard against the inappropriate disclosure of protected University data when using GenAI.”
Noting the three classes of University data: private, restricted and public, Hart said anything “considered to be the standard where those things are discussed, where those classifications are made,” is included via the University Data Risk Classification website link found in the document. This assists users in learning what is considered private or restricted data, and to “use those only with the tools that the University has approved.”
“Our data is not shared with others. It’s not made available to people outside the University community and those kinds of things,” he explained, noting the standard has been reviewed in various forums, most recently at the Information Technology Advisory Committee meeting, where it was endorsed.”
As a Pitt technology-based operating standard, Hart said it has the effect of policy “in the sense that it’s enforceable,” but is not a policy document that goes through the normal University policy-process channels, “if that makes sense.”
Robin Kear, Senate Council president, said because the document remains in draft stage, it’s too early for a committee vote.
“It’s just for us to be able to understand and provide any comment for you, since it is still a draft,” she said, noting she saw the document in an earlier stage and “made comments before. “But I think that it’s something that we need to consider and try to provide some guidance on.”
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.