Featured
The discussion was part of a dialogue put on by the Special Competitive Studies Project exploring innovation and national security topics.
Northeastern President Joseph E. Aoun’s 2017 book “Robot-Proof: Higher Education in the Age of Artificial Intelligence” lays out a blueprint for higher education as the world confronts unprecedented technological change.
In the seven years since its publication, there has been a particular development in the field of AI that Aoun says is changing the way universities, policymakers and the private sector envision that future.
“What has changed is something that we’re all aware of by now, and that is the LLMs [large language models],” he told an audience of business, technology and policy leaders in Cambridge, Massachusetts, on Friday. “AI didn’t start yesterday, it started 70 years ago.”
“However, with LLMs and generative models, we see more powerful tools — and they’re working by brute force,” Aoun said. “Those tools are very powerful, and at the same time, they have limitations.”
Friday’s discussion was part of the Ash Carter Exchange, a dialogue put on by the Special Competitive Studies Project exploring innovation and national security topics. The exchange honors the legacy of former Secretary of Defense Ashton Carter, an influential government official and academic.
Alongside other speakers, Aoun engaged in a fireside chat, titled “The Scientific Frontier,” anticipating developments in AI over the next five years. Aoun was interviewed by Amy Brand, director of MIT Press, publisher of Robot-Proof. The second edition of the book was released this fall.
What are the limitations of large language models? They “do not have common sense, they do not have logic and — more importantly — they do not understand context,” Aoun said.
Machines, Aoun said, cannot effectively transpose knowledge from one context to another — at least not for the foreseeable future.
It’s for those reasons that Aoun said he does not believe the so-called singularity, referring to a theoretical point in time in which technology becomes uncontrollable and unstoppable, is just 10 years away, as some in the AI space have suggested.
But Aoun said that the rise of generative AI — and the sheer pace of change — has underscored the importance of both experiential and lifelong learning, and the role educators play in shaping opportunities for the generation of students.
As ever-more-sophisticated generative AI tools go public, knowledge increasingly gets commodified, Aoun said.
“But experience is not,” he said. “Therefore, you need to rethink your educational mission along these lines.”
Aoun breaks down “the way universities deal with knowledge” in three ways: a desire to understand and explain the physical world; to understand and explain life; and to understand and explain social interactions, or society. Operating within these categories is, he said, “an understanding of the interaction between those three dimensions.”
But the arrival of sophisticated AI adds yet another dimension.
“The whole idea becomes how to integrate, in our educational mission, the fourth dimension, the fourth world, and help explain to our students the implications,” he said.
“The mission of the university now is to make people become robot proof,” he said, “to master humanics, and master the three literacies.”
As described in Aoun’s book, the three literacies include technological literacy, data literacy and human literacy. Together, they comprise “humanics,” an integration of the three literacies that allow students to take the knowledge gained from one domain and apply it to “a very different domain.”
Aoun added that it’s not enough to master the three literacies. “You have to put it into practice in an experiential setting: through co-ops, through internships.”
This is where Northeastern has stood out for years: experiential learning by way of its signature co-op program.
Asked about the global race to develop AI, Aoun spoke a bit about the relationship universities have to other sectors.
“We are very close in terms of the race,” Aoun said of China and the U.S. “Moving forward, we need to have the triangle involving universities, companies and the government on this national competition — and frankly, we don’t have a blueprint for that yet.
“Essentially, universities work well with businesses, with some limits; they work well with the government,” Aoun said. “But the opportunity I think is to put the three together, not only in terms of research, but having a national blueprint for how we can remain competitive at the human level.”
The Special Competitive Studies Project’s stated mission is to “ensure that America is positioned and organized to win the techno-economic competition between now and 2030, the critical window for shaping the future.”
Aoun recalled a conversation he had with Carter while lecturing in one of his classes: “He said to me … technology is unstoppable, but it’s shapeable.”
“There are two ways of shaping the advances of technology: one is to look at the technology itself; another is to look at the implication of technology on human beings,” Aoun said.