Skip to main content

AI Tutoring Has a Lot to Offer. But so Does Human Mentoring.


May 10, 2023

In education policy circles, there’s a lot of enthusiasm regarding the promise of AI-enabled tutoring. After all, a huge stumbling block for tutoring has been the limited number of affordable, reliable, and skilled tutors. That’s why ubiquitous AI could be such a game-changer.

But there’s reason to fear that the excitement of AI-enabled tutoring will distract us from some of the other, less fantastical possibilities offered by web-enabled mentoring. Look, I want to be clear. I am, and have long been, a believer in the promise of computer-assisted tutoring—especially when an AI tutor knows a student’s vocabulary or can inhabit fictional characters.

That said, it’s crucial to appreciate, as I explain in The Great School Rethink, that tutoring is mostly about conveying content and explaining concepts, while mentoring is beneficial precisely because kids also benefit from access to a broad array of role models and supportive adults. AI tutoring can help with the former, but not the latter.

And we can’t lose sight of that, especially because modern life compounds the “bowling alone” phenomenon first identified by Harvard professor Robert Putnam in the 1990s. Putnam noted, across lots of different measures, that Americans were joining fewer groups and spending more time in solo activities than they used to. Today, kids are enmeshed in fewer social networks, as they’re less likely than they once were to engage in things like church groups, the Boy Scouts, and 4-H clubs. As a result, they encounter fewer mentors, with implications for everything from college admissions to landing a job.

This all matters more for children with fewer advantages because, as Putnam has observed in his book Our Kids, networks of college-educated parents tend to include many more people with “social influence” (like politicians, CEOs, and professors) than do those of the less-educated. Celebrated economist Raj Chetty has found that connections to the affluent significantly increase the economic mobility of low-income children.

Mentoring has historically been limited by where people live and who happens to reside in that community. Low-income families are less likely to live down the block from or casually mingle with networked professionals. Digital technology can help change that, but by leveraging grown-ups rather than AI. For instance, in Who You Know, Julia Freeland Fisher and Daniel Fisher explore how schools can expand students’ access to relationships that might otherwise be out of reach, noting that students can now hear virtually from a range of adults about an array of jobs.

As Julia Freeland Fisher has put it, “Whom you know matters and what you know matters, but especially powerful is who knows you know what you know.” Recommendations from those with clout or connections carry added weight for youth, especially for kids with fewer advantages. In a world of active retirees, remote workers, and the self-employed, it’s easier to find adults with the flexibility, time, and interest to serve as mentors.

In practice, of course, it can be daunting to create a network of willing, reliable, and vetted mentors. Fortunately, there are organizations like CommunityShare or ImBlaze that can help with just this. Tucson-based CommunityShare, which exclusively serves Arizona schools, has been used by more than 11,000 teachers to locate community members who can address particular topics or lessons. ImBlaze, which works with more than 100 schools across the country, links students with local internship opportunities and provides a reporting system that allows employers and schools to monitor hours, submit required reports, and share essential information regarding openings and experiences.

As I observe in The Great School Rethink, “Connecting students to professionals, mentors, and the well-off can benefit all kids, of course, but especially those less likely to make those connections outside of school. While none of this is easy, schools today can readily stream experts and mentors into classrooms or advisories—something that would’ve been impractical even a decade ago.”

So, while we should welcome the possibilities of AI-enabled tutoring, a couple of big caveats are in order. We can’t let it crowd out the human dimension of mentoring. We shouldn’t focus so intently on the dazzling promise of AI that we overlook the way in which digital tools can foster mentoring and relationships. And we should be conscious of the downsides of encouraging socially isolated teens or tweens, who already spend six hours or more a day on devices, to spend more one-on-one time with digital devices.

Schools need to provide students with the kind of academic tutoring they need and also with the kinds of (in-person or virtual) mentoring that matter so much. This isn’t an either-or. And we shouldn’t let enthusiasm for AI cause it to become one.