Referencing the higher education treatise by Clark Kerr – “The Uses of the University” – Gerald Chan proposed several modern-day updates to the 1963 classic during a recent lecture at UC Berkeley.
Chan was invited to campus to give the first inaugural CDSS Dean’s Distinguished Lecture on February 4 to a packed auditorium in Sutardja Dai Hall. His lecture – titled “Rethinking Clark Kerr: The uses of the university in the age of generative AI” – reflected on Kerr’s contributions and considered how to prepare today’s students for a world enabled by AI technologies.
A scientist and venture capitalist in biotechnology, Chan has started over a dozen companies with intellectual property spun out from universities. His work has resulted in first-in-class drugs, diagnostics and medical devices approved by regulatory agencies for use in the clinic. In recent years, Chan has focused on using AI to develop digital biomarkers, digital diagnostics and therapeutics for dementia, autism and behavioral disorders.
Jennifer Chayes, dean of the College of Computing, Data Science, and Society (CDSS), introduced Chan and moderated a conversation with Chancellor Rich Lyons after the lecture.
Benefits to society, modern challenges for education
Chan reminded Berkeley faculty, staff and students in the audience that Kerr was ahead of his time in coining the term “multiversity” to describe the growing complexity of university aims and intended purposes, including their utilitarian value.
“Knowledge should become a driver of innovation to benefit society,” said Chan during his lecture.
An economist, Kerr served as the first Berkeley chancellor when the role was established in 1952. He went on to serve for 10 years as chancellor of the University of California, opening campuses in San Diego, Irvine and Santa Cruz and – with the UC Regents – transforming UC into a public university system.
Chan discussed Kerr’s contributions in the historical context of American universities, their close ties with federal and state government after World War II, and challenges related to social movements for civil rights and free speech.
Moving to present day concerns, Chan cited a July 2024 Gallup poll showing that only 36% of Americans have a great deal of confidence in higher education and 32% have little or no confidence. Reasons provided by survey respondents for a lack of confidence included the perception that colleges are pushing certain political agendas and not teaching the right kinds of skills. “If American higher education was a business, the strategist would say that customer satisfaction is abysmal,” he said.
Additionally, colleges and universities face financial pressures that have led some to cease operations in recent years. Mentioning a 2023 book by Brian Rosenberg, president emeritus of Macalester College, Chan suggested the title provided some indication of cause: “Whatever It Is, I’m Against It: Resistance to Change in Higher Education.”
AI as an education game changer
“Into this bleak state of affairs, enter generative AI, which will undoubtedly be both a game changer and an opportunity for every facet of society,” said Chan. “We will have to evolve our thinking on what constitutes an educated person and, with that, how will we educate.”
“As the internet is about content and knowledge, generative AI is about pedagogy,” he said. “If the internet had turned the availability of knowledge from scarcity to abundance for the everyday man, AI is turning pedagogy from one-size-fits-all to personalization.
Considering the use of traditional lectures now, Chan said that student behavior is revealing. “If you sat at the back of the lecture hall and looked at the screens of the students’ open laptops, it is clear most of them are parallel processing some other task while the professor is lecturing,” he said with a laugh.
Generative AI technology holds great promise toward providing personalized learning opportunities for students, Chan posited. He referenced a 2023 learning experiment at Harvard with an introductory physics course for life sciences majors. One group of students in the course was taught by an in-person instructor, and the other was taught by an AI tutor.
According to preliminary results in a preprint article by Gregory Kestin et al, “the students taught by the AI tutor learned faster, and the learning gains doubled that of the university teacher,” Chan said. “Moreover, the students felt more engaged and motivated.”
“The challenge for education is one of knitting together what technology is good at and what humans are good at,” he said. “Education is always a human and social activity. But we must reimagine the classroom, re-examine each aspect of the educational process, and redesign the outcome.”
Chan suggested offering students more experiential and project-based learning opportunities, such as internships with industry. He also proposed updating traditional notions on liberal arts education. “A liberal arts education is not the matter of subject or field of study,” he said. “It is a pedagogy with the aim of developing the learner's intellect.”
A plurality of purpose
After Chan’s lecture, Chayes opened the discussion by asking Lyons his thoughts about the widespread adoption of generative AI in education and research.
“We have to be vigilant about the impact of AI,” Lyons said. “If universities could be perfect at any one thing, one of the things I think they want to be perfect at is instilling the sense – or enlarging the existing sense – of curiosity. AI may allow us to get that little extra dose of payoff for what we're learning. I'm quite optimistic this sort of payoff will change enough that we'll have curiosity in an even stronger way. It may be a different kind.”
Chan stated that universities should consider a plurality of purpose and the challenges inherent therein for decision-making.
“Every student that comes to any school is a unique piece of work,” Chan said. “How can we accomplish these multiple purposes of civilizing them with a liberal arts education, and yet training them with the professional education so that they can go out and make a living? That decision is not entirely ours. It's very much the students we will turn out, some kids who are not the least bit civilized by a liberal arts education, and that's okay. It's up to them, but as a ‘multiversity,’ we offer all the resources.”
A question from the audience referenced recent news about the new DeepSeek model bringing to mind questions about public vs. proprietary information and open source vs. private AI models. What is the university's role in advancing open science, AI and the public good?
“Open source platforms raise the entire ecosystem,” said Chayes, “and then proprietary companies can be built on top of that. The investments that you're putting in, you'll get from those proprietary layers, but the platform itself will create ecosystems on which many, many companies can be germinated and can thrive.”
As an example, Chayes mentioned Apache Spark, an open-source data processing system originally developed at Berkeley and now used by thousands of companies and nonprofit organizations, with many additional applications developed.
“At Berkeley, there is really this mindset of questioning the status quo,” said Lyons. “There’s got to be a better way to do this. Berkeley has change in its ethos.”
