On Sept. 2, the first morning of fall classes, Harvard College Dean David Deming sent undergraduates an email saying he favored “actively encouraging AI usage where it can deepen learning.” He urged professors to shift toward assessments, such as proctored exams, in which students can’t turn to AI (that is, use it to cheat). He also suggested integrating the technology into writing courses rather than trying to police its use. Some horrified humanities professors pushed back . Writing itself, one argued, is part of learning to think. Using AI, another worried, might discourage students “from developing their own responses to works of art, music, literature, and philosophy.” For now, individual Harvard professors are still setting their own rules. (By contrast, the often fractious Harvard faculty recently agreed on a universal grading policy, limiting the number of A’s in each class.)

Welcome to the uncertainty (and even chaos) greeting today’s college freshmen, the class of 2030.

Higher education is already struggling with research funding cuts and political attacks from the Trump administration, an enrollment cliff following a birth decline that began in 2008, and growing doubts about whether a four-year degree is worth its rising cost.

Now colleges and universities are being forced to rethink something more fundamental: How should students learn in the AI age, and what sort of education will prepare them to thrive in that world? Choosing a college, already difficult, is even more complicated when schools are taking wildly different approaches to AI—and no one yet knows which one will succeed.

At Amherst College, President Michael Elliott is experimenting, too. Last fall, he told students in his literature course they could use AI on some assignments as long as they explained how they’d used it. “Which was also a learning experience for me, and I’m not sure what I’m actually going to do this fall,” he says. Amherst faculty are “all over the map” on AI, he adds, from returning to blue books and oral exams to embracing the technology. For now, the college has asked professors to clearly state on their syllabi what is—and isn’t—permitted. “We don’t yet have a grand strategy for how we’re going to do this,” he says.

Given that uncertainty, it makes sense to begin your college hunt with what we do know about educational success, and then ask some forward-looking AI questions, which we detail later in this story. To get you started, our new 2027 America’s Top Colleges list ranks the 500 best four-year colleges in the country. As we have since 2008, we focus entirely on student outcomes rather than test scores or reputation (which, to be frank, do affect outcomes). Our methodology includes, among other metrics, graduation rates, student debt, alumni earnings and achievement, and overall return on investment. We give extra credit to schools that produce strong outcomes for all students, including those from lower-income families.

The Massachusetts Institute of Technology captured the top spot for a second year running, boosted by financial aid so generous that students can graduate with little debt as they head off to graduate school or well-paying careers or their own startups. All eight Ivy League schools made the top 20, with five in the top 10. And for the first time, Vanderbilt University cracked the top 10, with a 94% graduation rate and low student debt. (The Nashville, Tennessee, school has made Forbes ’ New Ivies list of 20 schools employers love all three years we’ve published it.) Rounding out the top 10: Stanford University (No. 6), the University of California, Berkeley (No. 9)—again Forbes ’ top public university —and the California Institute of Technology (No. 7)—our new top small private college .

With families facing list prices at elite private schools that can exceed $100,000 a year for tuition, fees, room, board and personal basics, costs—and particularly net costs, after school financial aid and discounts—matter more than ever. At one extreme are 21 elite, wealthy private schools, all ranking in our top 50, that offer free tuition to families earning up to income thresholds as high as $250,000 a year. But with acceptance rates at these institutions as low as 3%, only a tiny share of students will qualify for those deals.

Families should also pay attention to Forbes ’ unique ranking of the 25 Colleges With The Highest Payoff. This eclectic group includes eight of those pricey schools with generous aid, along with nine City University of New York schools that keep tuition low while boosting graduates’ earning power. Yep, these are longtime commuter schools preparing students for careers on Wall Street and in education, criminal justice and tech.

Outcomes are measurable. What they can’t show is what the job market and higher education will look like four years from now. “The only thing we can count on is that things are going to change,” says Karen Brennan, the Timothy E. Wirth Professor of Practice in Learning Technologies at the Harvard Graduate School of Education. Colleges, she says, remain “a beautiful place to develop” the capacity to handle that change. “AI is not obviating that—it’s just giving us a new way to think about the importance and the necessity of adaptability.”

The AI playbook is being written and rewritten in real time. But forward-looking schools—and families—should ask the big questions now, then revisit the answers as evidence emerges about what really works. Here are some questions that students and their parents should consider.

Does the School Have Clear AI Rules?

Colleges may not come to the same answers, but they should at least be developing clear AI policies so students know what to expect—even if the rules vary from class to class.

Why does a school’s policy matter? Students are demanding it. “The student government has been asking to have fairly uniform guidelines and guidance,” says Benjamin Hermalin, vice chancellor and provost of the University of California, Berkeley, where professors set their own AI rules. Although it’s “not the Berkeley way” to dictate how teachers should handle a new technology, he says, “I think we will slowly evolve to having a little bit more standardization because the students really want it.” They want it, he adds, in part because they worry that classmates using AI to cheat will gain an advantage.

Harvard’s Deming raised the same concern in his email to undergraduates. Unenforceable AI bans, he wrote, create “a difficult dilemma for students who don’t want to cheat but also fear falling behind … their classmates.” Harvard (No. 3) plans to develop advisory guidelines this year.

The University of Michigan (No. 21) is trying to find a middle ground. On Sept. 3, it released a draft of universitywide guiding principles for AI in teaching and learning and invited the campus community to comment before they are finalized. The plans include “responsible, human-centric AI,” along with focusing on ethics, integrity, protecting privacy and more.

The best way to find out what AI use actually looks like at a university is to ask the students, advises high school history teacher Steve Fitzpatrick, a Forbes contributor and author of the Substack Teaching in the Age of AI . “You learn much more by asking the students questions because they will reveal things that may be off-script for the school,” he says. “Like, ‘How much are kids cheating here?’”

Are Professors Getting Help to Figure Out AI?

Students aren’t the only ones learning how to use the new technology. Prospective families should ask what a college is doing to help professors keep up. And look beyond one-time training sessions or a list of approved tools.

Brennan says faculty need two things: time to experiment and institutional support that helps them learn from what other professors are trying, both inside and outside the classroom. “It can feel really daunting if you don’t have the opportunity to get messy with tools or practice yourself,” she says. “Knowing what other people are doing, being able to share our experiments, our successes and our failures—that is what is going to help expand our pedagogical imaginations.” Students should be a part of that learning and experimentation, too. At Harvard, faculty can get individual coaching and attend workshops, while she says students have organized their own events and invited outside experts to campus.

On a campus visit, ask a professor for a specific example of how they’ve changed an assignment or exam because of AI. Ask students whether professors explain their reasoning behind their AI rules—and if students have a voice in determining them.

Will I Learn to Use AI No Matter My Major?

Students shouldn’t have to major in computer science to gain experience using AI. The University of Florida (ranked No. 34 overall and No. 9 among public universities), offers one model. After a $70 million gift from Nvidia cofounder and alumnus Chris Malachowsky, UF built what it calls an “AI University,” designed to send every graduate out the door with some grounding in artificial intelligence. It now offers more than 200 AI-designated courses across disciplines. Nearly 70% of graduating students have taken at least one, says Hans van Oostrom, director of its AI Academic Initiative. “Students of all majors need to know something about AI,” he says. “It’s not just computer science and engineering.”

But a long course list doesn’t mean students are learning to use the technology responsibly. Arizona State University (No. 110) is pushing aggressively into AI. “I think our philosophy quite simply has been to engage early and often,” says Anne Jones, ASU’s vice provost for undergraduate education. But Jones says simply having used ChatGPT, Claude or another generative AI tool doesn’t make a student AI literate. ASU expects students to understand how AI works, evaluate its biases and limitations and consider whether using it in a particular situation fits with their personal values. Just as importantly, students should learn how to use AI effectively within their own disciplines—something that will look very different for a chemistry major than for a screenwriter.

That means prospective students should look beyond whether a college offers an AI-related major or boasts about its development of an internal school chatbot. Jones suggests asking what the institution’s broader conversation around AI looks like, what concrete steps it has taken and how those decisions will affect a student in the field they want to study. “To be able to get both the institutional philosophy around ‘how are you engaging this,’ but also to get some concrete examples is what really gives people confidence that there’s some substance there,” she says.

Will I Develop Skills That Matter Alongside AI?

The need to develop “critical thinking” to succeed in the AI age is invoked so frequently that it risks sounding like a cliché. Ditto the need to develop skills of human interaction. College counselor Brennan Barnard suggests asking, “What are the ways that the school is preparing students for those durable, human skills?”

“These are the so-called soft skills, but they’re not soft. They’re actually incredibly difficult,” says Elliott of Amherst (No. 30 overall; No. 5 among small schools), who points to the ability to work with people from different backgrounds, persuade an audience and listen. “You learn them by being face to face in community with one another,” he says. “My advice to families would be to ask: Is the institution trying to use AI as a crutch rather than as a tool?”

Ask about specific courses geared to enhance these skills. Stanford University, for example, is expanding its required Civic, Liberal and Global Education (COLLEGE) program for freshmen to three quarters, from two, starting in fall 2027. Part of the general education curriculum, it asks students to consider “what ideas, scientific evidence and art can contribute to your personal development,” and exposes them to ideas beyond their chosen fields. “College is about more than just getting your first job,” says program director Dan Edelstein. “Of course, that’s important, but that’s not the full extent of what a college education can do. It’s really setting you up for life.”

Will This School Prepare Me For an Unpredictable Job Market?

There may be no such thing as an AI-proof major. Elliott is skeptical of anyone promising that a particular field is a pathway to economic success. “What you major in is going to matter even less in the future than it does in the present,” he says.

Barnard advises students to pursue their genuine interests while preserving “lots of optionality.” That might mean pursuing a double major, taking broad coursework or developing skills in research, writing and data analysis. Before choosing a college, ask how easy it is to switch majors, add a second one or try out different subjects without delaying graduation.

The career services office is also worth a visit. “In my view,” Barnard says, “it should be a stop on every college tour.” Ask there how the college is responding to changes in entry-level hiring and what services it provides to connect students with internships long before graduation.

At Drexel University (No. 147), more than 90% of students participate in its co-op program, President Antonio Merlo says. Students can complete up to three six-month paid jobs beginning sophomore year, moving between the classroom and the workplace. As employers cut entry-level positions, Merlo argues, “Our students have [already] done the entry level.”

Melissa Loble is chief academic officer at Instructure, the company behind Canvas—the online learning hub where professors post homework and grades and students turn in assignments. She suggests asking a college directly: “How do you use AI to better prepare me for what I do after I graduate?”

F amilies choosing among Forbes ’ Top Colleges understandably want some measure of certainty. Our ranking identifies institutions with strong records of graduating students, limiting debt, producing high-earning alumni and delivering a return on their investment. It can’t predict how AI will reshape college or the career prospects of the class of 2030.

Nor, Elliott argues, should families trust a college that pretends it knows the answer. “Institutions that are honest about the fact that we’re in a moment of uncertainty, to me, have an advantage over institutions that claim to have clarity,” he says. And learning to live with the discomfort of not knowing the answer is itself an important part of education. “Certainly,” Elliott says, “figuring out how to find pleasure in uncertainty is going to be a valuable life skill.”

Additional reporting by Asia Alexander.