🧠 AI Ed Wiki

📄 Research Article

Methodologies for Improving the Quality of AI Tutoring in K-12 Education

● high · 2026-08-13ai-tutoring, intelligent-tutoring, k-12, llm, personalized-learning, efficacy-study, engagement-metrics, edtech-platform

Synthesis: Udeshi et al. (2026), the team behind Khanmigo (Khan Academy's K-12 AI tutor, launched 2023), describe the metrics they use to measure AI tutoring quality and student engagement, along with the live experiments that have moved those metrics. Given that LLMs are opaque black boxes, they argue robust evaluation and live experimentation are essential. The paper highlights changes across models, prompting, personalization, and agents that improved tutoring outcomes. Accepted at AIED 2026, it connects to AI Tutoring, Intelligent Tutoring, and Research Methods AIED literatures.

Evaluation as the Engine of Improvement

Many AI tutors leverage large language models today. Because LLMs are opaque black boxes, robust evaluation and live experimentation to measure the impact of every change are essential. Khan Academy pioneered AI-powered tutoring for K-12 with the launch of Khanmigo (2023).

What They Measure and Change

The paper describes the metrics used to measure AI tutoring quality and student engagement, and the various experiments run to improve them. It highlights the changes that moved their metrics, including changes to models, prompting, personalization, and agents.

Position

This practitioner account from a major edtech platform grounds the AI Tutoring and Intelligent Tutoring literature in real, large-scale K-12 deployment evidence, complementing controlled Research Methods AIED research.

Connected Concepts

  • AI Tutoring
  • Intelligent Tutoring
  • K 12
  • LLM
  • Personalized Learning
  • Research Methods AIED
  • Engagement Metrics
  • Edtech Platform
  • Student Experience
  • Prompt Engineering

    • Connected Articles

  • Oatutor Open Source Adaptive Tutor 2023
  • GenAI Tutor Engagement Patterns
  • Measuring LLM Tutors Teach Vs Solve
  • Correct Answer Trap AI Tutor
  • From Answer Generators To Reasoning Facilitators AI Tutors
  • Access Not Enough AI Tutoring 2026
  • Tutoring Effectiveness Index
  • Deeptutor

    • Citation

    Udeshi, T., Khazenzon, A., Khan, K., Breen, N., Corwin, R. J., DiGiano, C., Weatherholtz, K., & Zaluski, M. (2026). Methodologies for improving the quality of AI tutoring in K-12 education. In Artificial Intelligence in Education (AIED 2026), LNCS vol. 16582. Springer. arXiv:2608.11259.