Stanford’s SCALE Initiative finds human-led tutoring well evidenced and AI-only tutoring unproven. For districts, that makes tutors and teachers the core of any technology plan.
Every few months, a new claim arrives that software will soon replace tutors, and eventually teachers. A calmer view comes from an August brief by the SCALE Initiative at Stanford University, “AI Tutoring is Not a Monolith: What We Actually Know,” written by Chayne Turano, Valeria Pihl, Chris Agnew, Lauren Ziegler, and Susanna Loeb.
Its core finding is simple. Human-led high-impact tutoring is well evidenced, with students gaining three to more than 15 months of learning across grade levels and subjects. Human-led tutoring supported by AI behind the scenes, for scheduling, feedback, and tutor training, has emerging evidence. AI tutoring with human support is emerging too. AI-only tutoring is rated unknown. The executive summary puts it plainly: “High-impact tutoring remains defined by live human-led instruction.”
Why the Human Part Matters
The brief describes human tutors as providing social-emotional support, troubleshooting, and redirection, which drives dosage and relationships. One provider says, “dosage is absolutely key,” and another, “humans are still at the heart of it.” The numbers explain why. In a study of 181,000 students on a supplemental math platform, only 5 percent reached the recommended 30 minutes per week and 41 percent never logged on. In AI-only settings, 40 to 47 percent of students never used the platform.
The strongest evidence in AI tutoring is the part with a human being in it.
AI as a Helper
The brief does not argue against AI. Students whose instructors used an AI system were 4 percentage points more likely to master lesson topics, and the benefit rose to 9 points for students of lower-rated tutors. That suggests AI may help newer tutors improve, which matters for districts hiring many of them.
The brief is also candid about what is unknown: emotional and developmental effects, long-term cognitive effects once a tool is removed, and data privacy. The research base lags behind how fast schools are adopting tools. A sensible district uses AI where it clearly saves time or improves human work, tests small, measures results, and does not cut human staffing on the strength of a promise.
A Voice on the Relationship Side
Tom Murphy, the founder of Sweethearts & Heroes, has reached a similar conclusion from the other direction: relationships carry the learning. A former All-American wrestler and UFC competitor who left a corporate career for full-time youth advocacy, Murphy started the organization in 2006 with Rick Yarosh. Its message is H.O.P.E., Hold On, Possibilities Exist, paired with action. On the home page, Murphy says, “We are training a generation of young people before whom the world will marvel.”
For educators, the organization lists an initiative called Become a Champion Teacher and a book titled 13 Pillows for Affective Teachers. Districts should review any program closely before adopting it, but the emphasis is the right one for hiring: the adult who builds the relationship is the most valuable person in the room.
This Is a Hiring Plan
Behind the brief’s findings sits a workforce. School-day tutoring with adequate dosage, three or more sessions a week totaling 90 minutes for at least 10 weeks, produced effects roughly twice those of out-of-school programs. In D.C. schools, students were 7 percent less likely to be absent on days with a tutoring session. Districts that want those gains need tutors, paraprofessionals, and teachers, scheduled inside the school day and supported with training.
Hire for the relationship. Ask candidates how they build trust with a struggling student. Use short demonstration sessions. Check references for how an applicant treats students. In postings, describe the mentors, training, and planning time new staff will receive.
Protect the People You Hire
Stable relationships matter, so turnover is costly. Districts can fund better pay and support by trimming overlap in recruiting costs. Many pay for several tools at once, including Frontline Education, SchoolSpring, EDJOIN, K12JobSpot, HigherEdJobs, and PowerSchool, through subscriptions, per-posting fees, or bundled modules. Count hires by channel, cancel overlap, move postings to a free option such as K12 Talent, and put the savings into teacher pay, tutor stipends, and mentor time.
Workload counts too. A Walton Family Foundation and Gallup survey reported that nearly two-thirds of educators who use AI said it freed time for the parts of the job they value, such as interacting with students, though the savings are self-reported. If AI really trims paperwork, districts should make sure the time goes back to students and to teachers’ own well-being, not into larger classes.
Tutor and paraprofessional roles can also be a pathway into teaching. With tuition help, mentors, and a clear route to certification, people who start as tutors can become the teachers a district needs most.
The full plan, with questions boards will ask and signals to watch, is in our blog on human tutors and AI.
Author Bio
Charlie Isham is the Founder and CEO of K12 Data, Inc., which operates K12 Data, College Leads, and K12 Talent. He has spent more than 15 years building education data infrastructure and has watched how schools, colleges, and the companies that serve them buy for over a decade. He is based in Encinitas, California.
