• 39 meta-analyses, 39 positive findings: An audit finds serious flaws in research on AI in education

    | The Hechinger Report

    Susanna Loeb, a Stanford professor of education who also joined the meta-analysis critique webinar, pointed out that her team has collected almost 900 studies on AI in education and determined that only 20 of them were of high enough quality to be included in a recent summary of the evidence for her AI Hub for Education. Her team has since identified 32 more studies and is soon to update that summary. But, she said, it’s still too early to conduct a meta-analysis that would be useful to school leaders or teachers. Very few of the high-quality studies take place in K-12 schools in the United States.

  • The Manufacturing of an Evidence Base for AI-in-Ed Through Flawed Meta-Analysis

    | MIT Teaching Systems Lab

    Some education stakeholders are familiar with the 2025 meta-analysis of Generative AI in education from the Nature Publishing Group that was retracted after receiving over 500 citations. Fewer people are aware that over a dozen similar meta-analyses have been published--many with the same serious issues as the retracted paper--and together they have over 2000 citations. Patrick O'Neill, faculty member at Ivy Tech Community College, has published a forensic review of 14 meta-analyses of AI in Education called "Presumed Effective: The Manufacturing of an Evidence Base for AI-in-Education Through Flawed Meta-Analysis." (https://osf.io/preprints/edarxiv/n29zd_v3) The paper compellingly argues that all 14 meta-analyses report positive associations between AI use and learning outcomes, have been published in leading journals, and have disqualifying flaws. Prof.

  • Congratulations to Chris Agnew, Named to District Administration’s Third Annual DA 100

    | District Administration

    Congratulations to Chris Agnew, Managing Director of the AI Hub for Education at SCALE Initiative, on being named to District Administration’s third annual DA 100 list.

    The DA 100 recognizes influential leaders and innovators shaping K–12 education. Whether based in a classroom, central office, or business headquarters, honorees share a common commitment: improving outcomes for students across the country and around the world.

  • AI Is Already in Latino Classrooms, The Federal Research System Built to Vet It Has Been Dismantled

    | The Latin Times

    The Alliance for Learning Innovation released its October 2026 report against that backdrop, citing a finding from Stanford's SCALE Initiative that underscores just how thin the research base actually is: of more than 800 academic papers on AI in K-12 education, only 20 produced strong causal evidence of learning impact. The gap between the volume of AI adoption and the volume of reliable evidence is enormous.

    "No state, district, or school should have to answer these questions alone, and the answers shouldn't depend on whether they have the resources to run their own pilots," ALI executive director Sara Schapiro wrote. The report's federal recommendations call on Congress and the administration to build a stronger evidence base for safe and effective AI use, improve AI-enabled tools, and construct the infrastructure needed to accelerate credible research and development across the country.

  • How SGOs Can Find Research-Backed Tutoring: A Guide for Using the Federal Scholarship Tax Credit

    | National Student Support Accelerator

    Starting January 1, 2027, people who give to scholarship granting organizations (SGOs) can get a federal tax credit. SGOs can use that money to pay for tutoring for students in public and private schools.

    Tutoring is a popular and research-backed way to help students learn. But not all tutoring works. This guide offers ways for taxpayers and SGOs to find tutoring programs built on the research.

    • Est. $26 billion (per year in donations by 2030)
    • 600-700 SGOs (expected to be running by 2030)
    • SGO-driven (SGOs, not states, choose what tutoring to fund)
  • Why Early Intervention and High-Impact Tutoring Can Be the Solution to Declining Literacy Rates

    | International Business Times

    Academically, high-impact tutoring is deemed the optimal model for early educational intervention. Stanford University's National Student Support Accelerator (NSSA) explains how this model can help a wide range of students, even those who are academically far behind.

    High-impact tutoring is a process that involves tutors, instruction, integrated learning, data use, equity, safety, and cohesion. This makes it expensive for most schools and families. As a result, far fewer students gain access to high impact tutoring than could benefit.

    This is where more effective reading interventions are required. At the center of this shift is Once and Matt Pasternack, co-founder and CEO. Once is an early-childhood reading platform that focuses on children from the ages of 3 to 7 years, attempting to dismantle the literacy crisis at its roots. The platform focuses on the 'science of reading' approach, empowering adults to provide in person reading tutoring to get the best results. The mission isn't to avoid human connection, but to enable it by equipping schools, teachers, and parents with the digital tools and structural frameworks that children need to excel academically on a day-to-day basis, as Pasternack explains.

  • The ‘Nonexistent’ Research on AI’s Benefits for Education

    | Inside Higher Ed

    Colleges are going all in on artificial intelligence. But it could be a long time before independent research emerges on how it should—or shouldn’t—be used to improve learning outcomes.

    “Can you get students to use AI tools consistently enough to even test whether or not they’re effective?” said Carly Robinson, director of research at Stanford University’s Systems Change Advancing Learning and Equity initiative, which tracks emerging research on AI and education. “The design has to be really deliberate just to get students to engage with these tools in the first place.”

    Although high-quality causal research is slow going, the limited findings so far support a measured approach to AI adoption in higher education.

    “There is increasing evidence that AI has the potential to benefit learning, but using it on its own without intentional design or guardrails probably reduces learning through cognitive offloading,” Robinson said. “There’s logic to the idea that AI can improve learning and potential for it to be transformational in education, but we’re a ways off from that.”

  • Less than half of states fund tutoring directly. Florida is one of them

    | Tampa Bay Times

    The big story: Florida is one of 24 states, including Washington D.C., that currently funds K-12 tutoring directly, according to a new review published today by Stanford University.

    The university’s third annual National Student Support Accelerator looks throughout the nation to analyze how states support high-impact tutoring and the policies and funding they have in place for it.

    High-impact tutoring is an immersive form of tutoring based on relationship building and individualized instruction known to improve education outcomes and even school attendance, according to lead author of the brief Kathy Bendheim.

  • Schools Ban AI For Students While Teachers Get It Free

    | Forbes

    There is also a growing body of evidence about what happens when AI is used directly with students.

    Stanford’s National Student Support Accelerator and its AI Hub for Education recently reviewed research on AI tutoring. One of the most interesting conclusions was that the strongest results were found where AI supported human tutors. AI-only tutoring was much less convincing.

    In some of the studies, engagement was a serious problem. In two randomized trials, 40% to 47% of students never used the AI platform at all, while those who did averaged only two to five minutes per week. There is an assumption behind quite a lot of education technology that giving a student access to a useful tool means they will naturally use it in a useful way. Often they do not.

  • OpenAI launches Learning Lab to build evidence base for AI in education

    | Ed Tech Innovation Hub (ETIH)

    The research network will bring together researchers, educators, students, and builders to study AI's effects on learning, teaching, assessment, and education systems

    OpenAI's Education team held the first Learning Lab convening at its San Francisco headquarters last week, bringing together researchers, educators, and builders from several countries.

    Researchers from the University of Tartu presented work connected to Estonia's national AI rollout and their study of how AI affects cognition over time. The University of Tartu and Stanford University are also working with OpenAI on new approaches to measuring learning outcomes.

    The meeting opened additional areas of collaboration, including research with Stanford's Accelerator for Learning, participation in the University of Oxford's AIEOU global community of practice, and work on challenges from Cornell University's National Tutoring Observatory.