Peer Tutoring Builds the Human Skills AI Cannot Replicate

Matt Wu does not dispute the educational usefulness of AI. It can tutor a student in calculus, write a semester-long study plan and make knowledge more accessible, personalized and intuitive. One Schoolhouse student, Wu says, even asked an AI to teach computer science in the manner of Gordon Ramsay on Kitchen Nightmares.

His argument is that access to knowledge does not itself supply the motivation or connection students often get from another person who keeps showing up. Schoolhouse, the nonprofit Wu leads, connects high-school students around the world for free peer tutoring over Zoom. The premise is simple: a student who has grasped a concept can help someone for whom it is still a stumbling block. The tutor reinforces their own learning and builds confidence; the learner receives support from someone who has recently worked through the same problem.

Schoolhouse reports 30,000 volunteer tutors and more than 100 million minutes of learning. Its capacity, in Wu’s account, is not based only on recruiting more adults. Learners who master material can become tutors themselves: one in five Schoolhouse tutors, he says, began as learners.

That reciprocal structure is the operating proposition behind Wu’s case for peer tutoring. Every phone or laptop can become, as he puts it, a portal to a trusted human; students can learn around the clock and across distance. It also produces a record of contribution, not merely of academic performance—evidence that a student can explain, collaborate, persist and help another person learn.

Wu contrasts that potential with his experience tutoring through Amphibious Achievement, a college program that taught Boston public-school students rowing and SAT preparation. The immediate activities were rowing and test prep, but its aim was to build what Wu calls an “unrelenting belief” in students’ own possibilities. He saw that belief surface in a rowing personal best, then a college acceptance letter, and eventually in a student’s changed sense of self. Yet the program was constrained by proximity, timing and the number of tutors available. Peer tutoring, Wu argues, can carry that kind of support beyond those limits.

The clearest learner outcome in Wu’s account is Carl McGrone, a 52-year-old from Ebenezer, Mississippi. Carl lost his father in high school, dropped out to support his family and concealed the fact that he had never graduated for the rest of his adult life. When he decided to pursue a high-school diploma at 52, algebra frightened him enough that he sought a tutor through Schoolhouse.

He was paired with Sachin Syal, a 14-year-old from Fremont, California who had joined Schoolhouse expecting to help students his own age. Instead, he met Carl. For more than a year, the two met every Sunday: an adult learner in Mississippi and a teenager in California, working on algebra.

A recording from the week before Carl’s exam captures the ordinary mechanics of the relationship. Sachin asks whether a step is “P, multiplied by two”; Carl corrects him—“It’s P divided by two”—and Sachin encourages him as they work it through. When Sachin asks how he expects to do on the actual test, Carl answers: “My confidence is beginning to kick in. I can’t get weak at algebra now.”

Carl scored highest in algebra, the subject he had once feared. Wu’s larger point is not simply that Carl found an effective explanation of the material. Carl was 14 when he left school—the same age Sachin was while tutoring him. Wu suggests that access to someone like Sachin at that earlier moment might have changed Carl’s trajectory.

The sessions therefore joined academic practice to a change in self-belief. Carl was learning algebra, Wu says, but he was also learning to believe in himself again.

Matt Wu presents Immanuel’s Schoolhouse work as an example of how institutions might recognize the human abilities peer tutoring requires. A first-generation student in rural Arkansas, Immanuel came to Schoolhouse looking for a way to explore his interest in teaching. He went on to lead more than 150 sessions in algebra, precalculus and SAT preparation, and his peers consistently rated him in the top 5 percent.

His Schoolhouse portfolio documented those contributions and accompanied his college applications. When admissions officers saw it, Wu says, they saw more than a letter grade: evidence that Immanuel could adapt to other students, help them with warmth, patience and engagement, and solve real problems with learners from 40 other countries.

Immanuel is now a sophomore at the University of Chicago and has told Wu that he wants to build an education startup of his own to pay the support forward. For Wu, care, collaboration and communication are not secondary traits. They are deeply human skills, and he says employers are seeking the same skills that cannot be automated.

Schoolhouse extended the peer model beyond academic subjects through Dialogues, a program that pairs students with differing perspectives to discuss issues such as immigration, gun control and abortion. There is no moderator and no script—just two students talking over Zoom.

In one conversation about free speech and hate speech, Joseph Z says that in China, free speech is not something people participate in. Bryan responds that if a politician uses hate speech, that should show at the voting booth. Afterward, Bryan says, “I’ve learned a lot today. I’ve had to consider things I never would have otherwise considered.”

The excerpts do not portray dialogue as an automatic route to consensus. In a discussion of income inequality, Akshaj Dadhich describes the sharp inequality he encountered growing up in India, while Wilson asks how business owners would feel about being required to give up equity to people they barely know. Laya says she generally does not talk about politics with friends. Giselli argues that humility matters in approaching these questions: “I may be wrong, you know, I need to do my research.”

In a discussion of science and technology regulation, Amrita N says she has “genuinely changed” her mind. Wu treats that willingness to listen and reconsider—not agreement as an assumed endpoint—as an educational outcome. Students arrive with conviction and practice taking another person’s experience seriously enough for it to shape their own perspective.

For Wu, education is not only the transfer of knowledge. It also involves learning to understand people who see the world differently, a process that depends on another human being on the other end of the exchange.