AI Across Disciplines event highlights the power of ‘breadth and connectivity’ at universities

In a packed auditorium at the recently opened Amy Gutmann Hall, Penn community members—deans, faculty, students, administrators, and beyond—gathered to formally launch the Penn AI Initiative and Penn’s role during this technological inflection point. 

Attendees were welcomed by Provost John L. Jackson Jr., who raised the “promise” of AI but also its “potential peril,” noting the importance for Penn to be a leader in AI research. He went on to acknowledge Penn’s unique position—and precedent—as the home of the first general purpose computer. 

“Penn has always been at the forefront of some of the newest technology,” Jackson said. “Our focus on interdisciplinary education and research means that we are, I think, uniquely positioned to both understand and advance artificial intelligence, which will touch every aspect of academic life and virtually every single human endeavor.”

Jackson recognized Penn’s undergraduate degree in AI, offered as the first in the Ivy League, through the School of Engineering & Applied Science, while also applauding the “spectacular new building” as a resource for the Penn community. Amy Gutmann Hall houses collaborations in data science between all 12 schools.

View large image

View large image

Jackson then introduced President J. Larry Jameson, who spoke about the significance of Penn’s—and universities’—contributions to AI. 

“Why should Penn, and In Principle and Practice, highlight AI as one of the pillars to focus on?” he questioned. 

Jameson went on to elaborate the University’s need to stay relevant to an advancement that will impact every sector of society, while also emphasizing the “breadth and connectivity” that make universities even better positioned than private industry to ask different kinds of questions and in a more interdisciplinary, dynamic way.

“We want to be relevant, but we’ve also got access to connectivity that just doesn’t exist [in industry],” Jameson said, comparing to how universities conduct medical research. “We’re in front of patients, we’ve got colleagues who do clinical trials, and there’s a bi-directional flow of information. So, this is a critical part of the interwoven nature at the University of Pennsylvania that is literally exemplified in this building, Amy Gutmann Hall.”

Following the screening of a brief video (below), providing an overview of AI at Penn and the launch of the Penn Advanced Research Computing Center (PARCC), Senior Vice Provost for Research Dawn Bonnell welcomed a group of panelists: Penn AI Council members Bhuvnesh Jain, the Annenberg Professor of Physics and Astronomy; Marylyn Ritchie, the Edward Rose and Elizabeth Kirk Rose Professor in the Perelman School of Medicine; René Vidal, a Penn Integrates Knowledge University Professor and the Rachleff University Professor in the School of Engineering and Applied Science; Duncan Watts, a Penn Integrates Knowledge University Professor; as well as Lynn Wu, associate professor of Operations, Information, and Decisions at the Wharton School, who filled in for AI Council member Eric Bradlow, the K.P. Chao Professor, also in Wharton.

In a wide-ranging discussion, panelists imagined many hard-to-fathom, yet within reach, possibilities of AI. Among them: finding a planet with life on it that surrounds a far-away star, AI-powered robots that could perform physical-world tasks, and AI models that could not only identify a tumor in a medical scan but explain their rationale for the diagnosis.

In medicine, said Ritchie, referencing the work of David L. Cohen University Professor of Pediatrics Kevin Johnson, there is enormous potential for AI to not only document changes in patients over time—think: observable changes in posture, gait, etc.—but also free up attention for physicians so that they can engage more with patients in the examining room, rather than take notes.

“If we could really change that interaction, I think healthcare interactions will be better for everyone,” she said. “And I think we’re already starting to see things like ambient listening; the Penn Medicine health system has already done some experiments with ambient listening and turning that audio into text that can go into clinician notes.”

Wu, meanwhile, described a “$1 trillion unlock” potential waiting to be captured through AI, but explained that know-how among businesses will be crucial. Already, she said, companies that have demonstrated effective use of AI have been rewarded with increased productivity and even added to the job market; companies that have misdirected and mismanaged their AI efforts, meanwhile, have struggled. 

With labor, she said, it’s important to see AI not as a substitute for workers but an opportunity to “re-engineer entire processes.”

“That’s where the value unlock happens,” she said.

View large image

Watts discussed skills and mindsets that could be useful for young researchers. He invoked Benjamin Franklin’s philosophy of addressing real-world problems through academia and suggested researchers have an approach of curiosity that assesses problems and determines which tools best address them, rather than making too many assumptions through a strict disciplinary or theoretical lens. 

“Building off of that, I think this initiative of the group we’re celebrating here, I think it can really help our students think like that and get them the tools they need from all the disciplines that we can bring to bear,” Watts said. 

Vidal, meanwhile, imagined a future where doctors will be able to use AI technology to search for every patient in the world who has had similar symptoms to what their patient is describing, and then leverage the data for a diagnosis. The key to accomplishing this, he explained, is ensuring physicians trust the algorithms and their probabilities—and can provide well-reasoned explanations for how they made their decisions.

“I think Penn is uniquely positioned to address that,” he said, pointing to the work of researchers like Aaron Roth, the Henry Salvatori Professor of Computer and Cognitive Science, who studies the design of ethical algorithms. 

Jain quipped that he is “a big fan of how AI has brought disciplines together—and I probably love that more than AI itself,” he said, to laughs from the audience. Students might be asking different questions, he said, but they’re starting to have a shared language around tools. 

“It’s been amazing how the tools of AI and the language of AI are shared by so many disciplines, not just science and engineering, but also social sciences and even literature and philosophy, to some extent,” he said, observing that students are “just fluidly moving across campus” in a way earlier generations of scholars might not have.

“This is an area of opportunity for the Penn AI Initiative. I think we can meet each segment of the community, of the different disciplines, where they are and boost their AI skills and performance.”