Smarter Support

AI tutoring can enhance high-quality math instruction

Research shows that artificial intelligence (AI) can support math learning through personalization, adaptive instruction, and data-driven insights, with positive effects on student achievement when paired with strong teacher instruction, explains NSBA's Jinghong Cai. 

July 20, 2026

DRAZEN/STOCK.ADOBE.COM

Demand for science, technology, engineering, and mathematics (STEM) workers continues to grow faster than for many other occupations. The U.S. Department of Labor projects that from 2024 to 2034, STEM employment will grow by about 8.1%, compared with 2.7% for non-STEM jobs. STEM careers also offer significantly higher wages, with median annual earnings around $104,000—more than double the roughly $48,000 median for non-STEM occupations. Mathematics proficiency is a key gateway to many of these high-demand careers. 

For school boards, this trend highlights the importance of prioritizing strong math instruction through policy decisions, resource allocation, and the adoption and effective implementation of high-quality math instructional materials (HQIM).

These resources are curricula designed to be meaningful and engaging for all students. They typically include grade-level content and tools that support teachers while reflecting local community needs (EdReports, 2021).

SMARTER SUPPORT FOR TEACHING AND LEARNING
Artificial intelligence (AI) refers to computer-based systems that can simulate aspects of human thinking and decision-making, including performing tasks or analyses using machine learning algorithms (Peterson, 2024). In the 2024-25 school year, 60% of K-12 public school teachers reported using AI tools in their work (Gallup, 2025). Math teachers described using AI to create problem sets, practice questions, word problems, and problems with real-world applications (RAND, 2025).

For students, AI tutoring in math education has advanced rapidly, with its use becoming more widespread and impactful over time. 

  • In the 1990s, Carnegie Mellon University developed early AI-powered math tutoring systems designed to match students with interactive digital tutors.
  • By 2016, AI tools began to be integrated into learning management systems, allowing schools to use data analytics to support student engagement and performance (Wang, 2025).
  • Since 2023, the emergence of large language models (LLMs) has accelerated the use of generative AI (GenAI) in classrooms, expanding capabilities beyond answer-checking to supporting complex problem-solving and conceptual learning (Sawyer, 2024).

Research shows that AI can support math learning through personalization, adaptive instruction, and data-driven insights, with positive effects on student achievement when paired with strong teacher instruction. However, studies also highlight important limitations, including accuracy concerns, age-appropriateness issues, and the risk of overreliance without strong teacher oversight. Overall, AI is most effective when used to support —not replace—high-quality teaching and instructional materials.

WHY SCHOOL LEADERS SHOULD CONSIDER AI-POWERED MATH TUTORING

Since the 1990s, researchers have used cognitive modeling to better understand how students learn mathematics and to design more effective curricula—both print and digital (Carnegie Mellon University, 2007). This work not only has advanced knowledge of mathematical cognition but also produced instructional approaches that have proven effective in real classrooms. Evidence from Carnegie Mellon University shows that millions of U.S. middle and high school students have improved their math learning through AI-based cognitive tutoring systems.

Research findings consistently demonstrate impact:

  • An intelligent tutoring system aligned to an algebra curriculum—focused on real-world problem solving and computational tools—produced significant gains among ninth-grade students in three Pittsburgh high schools. On average, the 470 students in the program outperformed their peers by 15% on standardized tests and by 100% on assessments aligned to course objectives.
  • A longitudinal study shows that more than 1,500 middle school students have used The Decimal Point learning game and its accompanying curriculum materials. Set in an amusement park, this game-based tool has been associated with measurable gains in students’ understanding of decimals and decimal operations.

The effectiveness of AI tutoring tools points to a promising solution to the persistent gap in access to high-quality, engaging learning opportunities (AIED, 2023). While high-dosage tutoring can deliver strong results, many students—especially those from economically disadvantaged backgrounds—lack access to well-trained tutors.

AI offers a practical way to extend the reach of human expertise, scaling tutoring support while maintaining instructional quality. By augmenting—not replacing—teachers, AI makes effective tutoring more accessible to more students at a lower cost.

The evidence suggests that AI-powered math tutoring delivers both proven effectiveness and a scalable pathway to expand access—positioning it as a practical strategy to improve outcomes while advancing equity.

HOW TO LEVERAGE AI TUTORING IN HQIM

There is a growing trend to integrate age-appropriate AI tutoring tools into HQIM. This is not simply a technology decision—it is a system design choice. When thoughtfully embedded, AI tutoring can strengthen systemic coherence by aligning what students learn, how teachers teach, and how progress is monitored.

To be effective, AI tutoring should be tightly aligned with HQIM—reinforcing grade-level content, instructional routines, and assessment expectations—rather than functioning as a standalone intervention. In 2025, the National Council of Supervisors of Mathematics (NCSM) published "Educational Technology & AI Guidance for Math Leaders." The guidance encourages school leaders to ask several key questions when considering technology tools: 

  • Does the tool deepen students’ mathematical thinking? 
  • Does it align with what we know about how students learn mathematics?
  • Does it strengthen teachers’ professional judgment or replace it?

With these questions in mind, school leaders should focus on the following criteria to guide decision-making.

  • Align first. AI tutoring must match HQIM—same scope, sequence, and pedagogy—so learning stays coherent across instruction, practice, and intervention. 
  • Empower teachers. Use AI to inform, not replace, teacher judgment with real-time insights on student thinking, progress, and misconceptions. 
  • Ensure access. Scale impact by guaranteeing devices, connectivity, and supports—especially for underserved students. 
  • Lead with governance. Set clear expectations for alignment, data privacy, and instructional quality; evaluate tools based on learning impact, not features.

In summary, when integrated effectively, AI tutoring becomes part of a coherent instructional system. It amplifies the impact of HQIM, strengthens teacher capacity, and improves student outcomes.

Jinghong Cai, Ph.D. is the senior research analyst of NSBA's Center for Public Education.