Intelligence Report

MIT research effort scrutinizes AI's effects on students, voters and cognition

A $5.5 million MIT initiative is funding studies of how generative AI shapes learning, political information and human relationships.

INSIGHTEX Editorial Desk

2 min read

MIT research effort scrutinizes AI's effects on students, voters and cognitionIntelligence Report

What happened

MIT students are visiting office hours less often and spending less time on problem sets and with classmates, according to a recent Institute report on AI and education, MIT News reported. Those findings have deepened faculty concern about generative AI. The MIT Human Insight Collaborative, known as MITHIC, has made the technology's effects on how people think, learn and relate to one another a central focus of its funding.

Since its 2024 launch, MITHIC has awarded $5.5 million to 67 projects across 31 MIT units, and roughly a third of them address AI, society and humanity. In one, Chara Podimata, an assistant professor at MIT Sloan, is working with political scientists Adam Berinsky and Charles Stewart III to examine large language models as sources of political knowledge in the run-up to the 2026 U.S. midterm elections.

That project found that Alaska voters who used Anthropic's Claude to research the midterms may have been given incomplete information about candidates and their positions. Other funded work tracks how AI affects thinking over both short and long time spans. In a new course on humane user experience design, anthropology professor Graham Jones said, students are finding that a humanlike system can conflict with one that is good for people, since users can form dangerous dependencies.

Context

MITHIC began in 2024 as a strategic initiative at MIT and is now in its third year. Earlier this year it co-organized an on-campus AI and Society Forum with partners such as the Institute's Social and Ethical Responsibilities of Computing program, as well as MIT's Generative AI Impact Consortium, covering jobs, democracy and AI. Co-lead Seth Mnookin said the most important questions about AI cannot be answered by technologists alone.

SourceMIT News

The briefing

The week in tech, in one email.

The stories worth your time, every week. Unsubscribe anytime.