At the recent AI and Society Conference at MIT, experts from across the Institute discussed the potential benefits and risks of technological innovation in the workforce, the nature of work, public discourse, election management, and other topics.
The event featured individual research presentations and panel discussions, as well as a musical performance exploring the use of generative artificial intelligence in the arts.
The forum is jointly organized by the School of Humanities, Arts, and Social Sciences (SHASS) and the Social and Ethical Responsibilities of Computing (SERC). It was launched in partnership with two MIT strategic initiatives: the MIT Generative AI Impact Consortium (MGAIC) and the MIT Human Insight Collaborative (MITHIC).
Agustín Rayo, Kenan Sahin Dean of SHASS, and Dan Huttenlocher, director of the MIT Schwarzman College of Computing, gave opening remarks.
Rayo said bringing together academics from across MIT was intentional because understanding the impact of AI requires expertise from academics across the Institute.
“Attention to the social consequences of AI is not a departure from MIT’s work; it is a way to ensure that our technological leadership has a greater impact,” Rayo said.
Huttenlocher added that computing and the rapid growth of AI make it more important to support interdisciplinary research and dialogue.
“Understanding where AI excels and where it falls short is important not only for unlocking its benefits, but also for avoiding critical mistakes, overreliance, and unintended consequences,” Huttenlocher said.
Jobs and AI
Held in Tull Concert Hall in MIT’s Linde Music Building, the May 12 forum opened with a keynote presentation from economist David Autor, Daniel (1972) and Professor Gail Rubinfeld in MIT’s Department of Economics. Autor challenged the common narrative that AI will simply eliminate jobs by suggesting that the technology’s impact depends on how it affects the scarcity and value of human expertise.
“When I think about how technology interacts with the labor force, I think about it in terms of how it changes the scarcity of technology, whether it’s making it valuable or making it more of a commodity,” he said.
Autor said what matters is whether automation removes routine support jobs or removes specialist jobs. He pointed out that AI will likely create new specialized jobs, which require effective policies regarding employee training, wage insurance, and more capital ownership.
A panel discussion followed, featuring experts from MIT discussing how work is changing and what it means for society.
Daniela Rus, MIT Panasonic Professor of Computer Science and director of the Computer Science and Artificial Intelligence Laboratory (CSAIL), expressed excitement about the ways AI can improve the workplace.
“I would like to see a robot as your friend and assistant, as a person who watches you and sees how he can help you as a person who can work at a high level,” he said.
Still, Rus said, human judgment is still important in decision-making.
“We could really think about working with AI tools, but the human role as a decision maker, as a person with good judgment, as a person who decides the next step, whatever it may be, remains very important,” he said.
David Mindell, professor of Aeronautics and Astronautics and the Dibner Professor of History of Engineering and Manufacturing in the Program in Science, Technology, and Society, says that the nature of work has changed over the years, but “what matters is new work.”
“We have to support individuals, the economy, technology, to keep creating new jobs,” he said. “It’s very important that we give young people the tools and let them do what they think is creative and show us what the new work will be.”
Panelists also addressed the need to maintain safety standards, while exploring ways to achieve efficiencies. Mindell used the example of cargo planes that require six pilots because of the length of the plane.
“We don’t know how to reduce that number from six to five, significantly reduce it to two, one, or zero. There is a lot of money behind solving that problem, but there is also a very rich plan that has emerged to make those plans safe,” he said.
Sendhil Mullainathan, a two-time Peter de Florez Professor in MIT’s Departments of Economics and Electrical Engineering and Computer Science (EECS), explained the idea of the use of AI and growth that provides productivity improvements, but he also cautioned, “I think it is very important to separate productivity gains from things that create long-term growth.”
Either way, Mullainathan said, it’s clear we’re entering an era of great diversity in terms of AI’s impact on the workforce.
“If you said, ‘how will organizations reorganize?’ I don’t know. But will there be more restructuring? It’s hard to believe that there won’t be more restructuring. And in a way, if we know that we are entering an era of great diversity, that in itself is incredibly instructive,” he said.
Democracy and AI
The second session of the day focused on AI technology and its impact on democracy.
Chara Podimata, Class of 1942 Career Development Assistant and assistant professor of career research and statistics at the MIT Sloan School of Management, presented her research examining large-scale linguistic models of bias in electoral information.
“Algorithms are deciding a lot of things about our lives right now,” he said. “Regarding chatbots and election information, if I take two people and they interact with the same chatbot … how is the chatbot going to respond? How is it going to customize the information that it gives these people?”
A longitudinal survey of 12 major models during the 2024 US presidential election season found responses varied widely based on demographics and political leanings. His research team is now working on a new study on the 2026 US midterm elections, using a redesigned survey with the opinions of political science experts.
During the panel discussion, experts raised concerns about the possibility of AI to destroy democratic norms and processes, but also explored the possible positive effects.
Bailey Flanigan, Theodore T. Miller (1922) Professor of Career Development in the Department of Political Science, who holds a position in the MIT Schwarzman College of Computing that shares EECS, said he doubts how others use AI as a tool that can make people reach decisions or consensus very quickly.
“And there is a reason to think that this is good because it works very well. It is simple. But it loses many of these aspects of the democratic process which are the habits of how we meet and make decisions,” he said. “And I think it’s a mistake to forget about that when we start thinking about automation.”
Charles Stewart III, Kenan Sahin (1963) Distinguished Professor of Political Science and founding director of the MIT Election Data and Science Lab, said another challenge is that government agencies are not changing at the same rate as technology.
Stewart said his main concern is that AI could lead to chaos during and after the election.
“And if things go wrong, they can go very, very wrong. If the election is contested, that can lead to violence,” Stewart said.
“We have already seen in the age of low technology election results being misused. My concern is what I will end up with on this coming Election Day, and next Wednesday, and if AI has helped to cause irreparable disruption in the electoral process,” he added.
Lily Tsai, Ford Professor of Political Science and director and founder of the MIT Governance Lab (MIT GOV/LAB), said that in many ways, AI opposes the democratic process and commitment necessary for a healthy democracy.
“It’s really important not only in terms of design principles, but the responsibilities of designers to be familiar with the values and principles that reflect what democracy is based on: agency, political equality, mutual respect, inclusion, and autonomy,” said Tsai.
Tsai also noted that her research has shown that some people are more comfortable interacting with machines. He described a “Socratic conversational chatbot” his team designed to ask people to reveal the reasoning behind their beliefs and positions.
“And that, interestingly, seems to reduce their policy position,” Tsai said. “So there are perfect examples of ways AI can have a positive impact on democracy. But it’s really about designing with the right principles and evaluating them carefully.”
