SDI Productions/E+/Getty Images
As artificial intelligence rushes onto our campuses, tutoring has become a telling test case for how easily we might trade human connection for apparent technological convenience. The appeal is obvious: AI tutoring promises scalable, individualized support for every student, available at any hour. On the surface, it looks like the safest and most helpful application of AI in higher education. Yet, beneath that promise lie risks that go to the very heart of how students learn, connect and ultimately thrive in college.
Too often, tutoring is imagined as a kind of academic side service—someone who can walk a student through calculus problems or proofread a paper. But that has never been its true purpose. Tutoring is fundamentally relational. It is a relationship in which students can admit uncertainty, ask vulnerable questions and feel seen by someone who has walked a similar path. Peer tutors not only explain concepts, but also model persistence, normalize struggle and help students find resources across the campus. This mix of academic and developmental support is what makes tutoring such a powerful equity practice.
That is why the research is so consistent: Students who use tutoring are more likely to persist and graduate. The benefit cannot be reduced to “content mastery.” It comes from the sense of confidence, belonging and trust that grows when students connect with someone who believes in them.
I offer this perspective not as a technophobe, but as someone who has embraced AI’s potential. As a teaching and learning center director, I work daily with faculty to explore how AI can empower teaching and learning. I am excited about those possibilities. I have seen how AI can spark creativity, support inclusive pedagogy and give faculty new tools to engage students. I am genuinely an enthusiast for its promise.
But as a scholar of peer tutoring—and as someone whose professional path has been deeply shaped by it—I am alarmed by the widespread enthusiasm for AI as a substitute. This framing misunderstands why tutoring is effective in the first place. Tutoring works because of person-to-person, student-to-student connection: the trust and solidarity that grow when learners support one another. An algorithm, no matter how sophisticated, cannot replicate that relationship.
AI has its strengths. It can generate practice problems, break down concepts into steps, and answer questions at 2:00 a.m. For many students, that accessibility is useful. But efficiency is not the same as connection. While it can emulate motivational speech, AI cannot sense the discouragement behind a pause, or share a story of failure to normalize struggle. It cannot walk a student to a counseling office when stress spills into academics. And it cannot offer the solidarity of a peer saying, “I failed my first exam too—let’s figure this out together.” Students know the difference. One is scripted encouragement; the other is human belief.
The danger is that institutions, under pressure to cut costs and scale services, will see AI tutoring as a solution too good to pass up. It is framed as consistent, endlessly available and cheaper than staffing dozens of student tutors. The apparent cost savings are appealing to administrators trying to stretch budgets. But those calculations ignore the hidden costs of replacing peer tutoring with AI.
When we do so, we lose not one, but two well-documented benefits. The first is the benefit to student learners: Tutoring’s highly relational pedagogy builds confidence, persistence and belonging—outcomes that cannot be simulated by a chatbot. The second, less often discussed but equally important, is the benefit to the tutors themselves. Serving as a peer tutor is one of the most powerful pre-professional experiences undergraduates can have. It allows them to practice communication, empathy and facilitation. It helps them begin to see themselves not just as students, but as educators. It bridges their identity as learners with their emerging identity as professionals. When tutoring is outsourced to AI, we rob both groups of students—the learners and the tutors—of experiences that shape their academic and professional futures.
What makes this moment especially perilous is how quietly it is unfolding. Many AI tutoring platforms are being adopted through procurement processes that treat them as technology upgrades, not pedagogical changes. Contracts are signed and platforms embedded before faculty or learning-center staff are meaningfully consulted. By then, the definition of tutoring—as something transactional that AI can replicate—has already been baked in.
Higher education does not have to accept this narrowing of vision. There are steps we can take now. Institutions can reaffirm that tutoring is developmental and relational, not merely transactional. They can slow down adoption and bring faculty, staff and students into the conversation before signing contracts. They can invest in peer tutoring as essential infrastructure for equity and retention, rather than treating it as an expendable add-on. And they can frame AI explicitly as a supplement, not a substitute—as a study partner or practice tool that extends, but never replaces, the work of human tutors.
The question of AI tutoring is about more than technology. It is about what we value in higher education. Will we let convenience redefine support in ways that strip it of its human core? Or will we remember that the deepest learning happens in relationship—that students succeed when they feel connected, seen and encouraged by others who believe in them?
Tutoring is one of the most humanizing practices we have. It affirms that education is not only about mastering content, but about growing into one’s own capacity as a thinker, a learner and a member of a community. If we let AI replace that, we will have chosen efficiency over empathy and convenience over connection. And in doing so, we will have lost far more than we gained.
