MIT Plans $1-Billion Project to Develop Artificial Intelligence — and to Tackle Challenges the Technology Will Create

The Massachusetts Institute of Technology announced on Monday that it would spend $1 billion on a new college within MIT to study artificial intelligence.

Artificial intelligence, or machine learning, involves the development of software that, just like the human brain, takes in data, weighs it, makes a decision, and, often, takes action, all without human intervention. Commonly known as AI, it is already revolutionizing such fields as customer-service calls and transportation, though it also carries concerns over the implications of machines’ taking the place of human beings.

In addition to furthering the practical progress of AI, the institute’s project will support the search for solutions to two other daunting challenges: how to handle the ethical and philosophical implications of AI for the societies it will transform, and how to break down institutional silos in academe.

The commitment to the new college is, MIT says, the largest expenditure on AI in American higher education. Stephen A. Schwarzman, co-founder and chief executive of Blackstone, a private investment firm, donated $350 million to MIT to seed the new college, which will be named in his honor. An additional $300 million has already been raised from other philanthropic sources.

The college will hold its first classes next fall, with a new building scheduled to open in 2022.

MIT plans to hire 50 faculty members for the new college, and they won’t all be computer scientists, or even situated completely at the college. About half will hold joint appointments with other colleges. For example, the institute might hire a scholar with a background in political science or philosophy, but he or she would also be “conversant and strong in these cutting-edge computational tools,” said Martin A. Schmidt, the provost.

True Interdisciplinary Work

MIT hopes to accomplish several things with the joint appointments, Schmidt said. For one, having a cross-disciplinary faculty will allow professors to apply AI to other disciplines, and then bring what they learn from that work back to the college to help improve AI. For another, connecting AI to other disciplines and vice versa will help break down some of the silos that prevent truly interdisciplinary thinking and research.

Generally speaking, Schmidt said, faculty members who hold joint appointments now too often get caught in the breach between the two disciplines, and professors hired to consider the impact of technology sometimes end up isolated from the technologists.

“We’re going to need to rewire the promotion and tenure process,” he said, to recruit and support faculty members who can help the college pull off its ambitious goals.

Andrew McAfee, a principal research scientist at MIT’s Sloan School of Management who studies information technology’s impact on business, said that true interdisciplinary work is rare in academe. But he’s encouraged by the possibilities for the Schwarzman College of Computing because AI has already inspired so much interdisciplinary discussion.

“As these questions around AI capability and AI progress have deepened, a truly cross-disciplinary, a truly diverse set of people at MIT and elsewhere have naturally started coming together and getting around the same table and trying to learn from each other and talk about these things,” he said.

AI would not represent the first time a new discipline required colleges and universities to rethink their structure and approach. Computer science itself was once squirreled away inside mathematics departments, said Jerry Kaplan, a tech entrepreneur and a lecturer and research affiliate at Stanford University.

Biotechnology, to name another example, not only has created new collaborations across established disciplines, but also has raised questions about what kind of research should be restricted, and the potential benefits and dangers of altering genetic code — concerns similar to some of the dilemmas facing AI researchers.

Considering how transformational AI could be for so many aspects of life, Kaplan said, “the need to integrate with lots of other disciplines is very significant.”

‘Classic Shiny Object’

Along with academe’s silo problems, MIT’s new college is designed to tackle the bigger questions about the advent of AI. It’s baked into the college’s reason for being.

In a written statement, Schwarzman said that society faces “fundamental questions about how to ensure that technological advancements benefit all — especially those most vulnerable to the radical changes AI will inevitably bring to the nature of the work force.”

McAfee believes that, like other powerful technology innovations, AI promises to make society more prosperous. But, he added, the distribution of that prosperity might change “in ways that we might not like, or in ways that might not strike people as fair or in keeping with the bargain they signed up for.”

If a certain piece of software could automate some critical function and put tens of thousands — or hundreds of thousands — out of work, he said, “there are policy questions that come up, there are really open economic questions that come up, there are ethical questions that come up.”

AI is “a classic shiny object,” said Malcolm Brown, director of learning initiatives at Educause, a nonprofit organization concerned with information technology in higher education. Many have rushed to embrace a cool, new technology without thinking too deeply about what it might portend. There are huge issues at stake, Brown said, and he’s glad to hear that MIT will be grappling with them.

As more companies develop AI technology, it falls to MIT and other universities to do the work of considering the broader implications of AI. “To be perfectly honest, nobody was thinking about them before,” said Daniel N. Rockmore, associate dean of sciences at Dartmouth College. “The companies didn’t care.”

If the Schwarzman College of Computing devotes substantial faculty resources to pondering, and teaching, the ethical dilemmas involved in the technology, he added, “maybe more of these future engineers and start-ups one day are going to remember their lessons, and that’s going to be part of their pattern of thought as they spin up new ideas and send them out into the world.”

Lee Gardner writes about the management of colleges and universities, higher-education marketing, and other topics. Follow him on Twitter @_lee_g, or email him at lee.gardner@chronicle.com.