Artificial intelligence can weaken the habits and institutions democracy depends on, but it may also help reinforce them. That was the central tension in Harvard legal scholar Martha Minow’s address at MIT’s fall Compton Lecture, where she examined the civic foundations of democracy in the United States.
Four foundations under pressure
Minow organized her remarks around four ingredients: education, reliable and accessible information, security from violence, and fairness. She described a difficult climate for democracy and constitutionalism, citing violations of laws and norms, pressure on free speech, reduced congressional authority, disregard for legal rulings, and the vilification of people in society.
Recent Gallup polling, she observed, found that 77 percent of people believe the founders of the U.S. would be disappointed by the country’s current condition. Minow said many people are questioning whether the nation’s system of government and rule of law, ignited 250 years ago, will endure.
Her account linked these concerns rather than treating them as separate problems. Education helps people assess civic questions; accessible information gives them material to assess; safety and fairness shape whether they can participate. Weakness in one area can make the others harder to sustain.
AI’s risks in education
Education received particular attention. Minow pointed to longstanding local and regional disparities in school funding and said schools have become significantly more segregated in economic and ethnic terms over the last three decades. She warned that AI could further erode independent thinking at a time when civic understanding matters greatly.
Some AI platforms and interfaces, she argued, can encourage passivity. Minow raised concerns that leading models, including sycophantic features, could interfere with young people’s ability to think, remember, and engage in critical thinking. The issue, in her view, is not only what technology can produce, but what people may stop practicing when it does too much for them.
She called for people to learn how to use and regulate AI, and to understand the costs of using it without thought. AI’s effects on learning are not fixed, however. Tools could be designed to act more like tutors, offering prompts, questions, and hints that help students work through ideas instead of replacing their work.
Minow said better tools and more meaningful, cheating-resistant assessments could help deepen curiosity and human activity. That approach would make the technology a support for learning rather than a substitute for the effort learning requires.
Information, safety, and fairness
Minow also connected democracy’s information problems to changes in media. She said modern media technology has contributed to a 75 percent reduction in journalists in recent decades, alongside increased corruption in places without local media. Her proposal for civic-minded technologists was to help build a widely used “information commons,” or multiple forms of one, where people could find more reliable information.
The idea speaks to a basic democratic need: people need access to information they can use to understand public issues. Minow’s remarks did not frame technology as a complete answer. Instead, she urged people to apply technical work toward shared civic needs, including dependable information and stronger community relationships.
On security, she warned about credible threats of violence against judges, describing the situation as unacceptable. She also discussed several dimensions of societal fairness. Together, these concerns point to conditions that affect whether public institutions and people can function without intimidation and exclusion.
Building the conditions for participation
Minow’s assessment was serious, but she also emphasized room for action. She said new forms of reliable information, education, and community relations could be created, and argued that people at MIT and elsewhere could devote their skills to practical solutions.
During the question-and-answer session, Minow said she worried about losing faith in reason, testing ideas, and rejecting myths when evidence contradicts them. She connected that concern to disinformation and the decline of critical thinking in schools, as well as to the rights, tolerance, and broader participation she associated with the Enlightenment.
Her challenge was both civic and technical: build AI that strengthens people’s ability to learn and judge, and help create institutions that make reliable information and participation possible. As Minow put it, “We have to actually invent the world that we want to live in.”