Research & Development: Khanmigo: Artificial Intelligence-Powered Tutoring and Teaching Assistance — Khan Academy

Citation
APA
Learn & Work Ecosystem Library. (2026). Research & Development: Khanmigo: Artificial Intelligence-Powered Tutoring and Teaching Assistance — Khan Academy. Retrieved 19 November 2026, from https://learnworkecosystemlibrary.com/initiatives/research-development-khanmigo-artificial-intelligence-powered-tutoring-and-teaching-assistance-khan-academy/
MLA
"Research & Development: Khanmigo: Artificial Intelligence-Powered Tutoring and Teaching Assistance — Khan Academy." Learn & Work Ecosystem Library, 18 November 2026, https://learnworkecosystemlibrary.com/initiatives/research-development-khanmigo-artificial-intelligence-powered-tutoring-and-teaching-assistance-khan-academy/. Accessed 19-11-2026.
Chicago Footnote or Endnote
"Research & Development: Khanmigo: Artificial Intelligence-Powered Tutoring and Teaching Assistance — Khan Academy," Learn & Work Ecosystem Library. 19-11-2026, https://learnworkecosystemlibrary.com/initiatives/research-development-khanmigo-artificial-intelligence-powered-tutoring-and-teaching-assistance-khan-academy/.
Chicago only requires the accessed date in a citation if no publication date is listed for the source. While many of the Library's entries include a 'last updated' timestamp, some do not, and for these you should include your accessed date.
Chicago Bibliography
"Research & Development: Khanmigo: Artificial Intelligence-Powered Tutoring and Teaching Assistance — Khan Academy." Learn & Work Ecosystem Library. 19-11-2026. https://learnworkecosystemlibrary.com/initiatives/research-development-khanmigo-artificial-intelligence-powered-tutoring-and-teaching-assistance-khan-academy/.

Subscription or Fee May Be Required to Access the Information on the External Link

Overview

Launched in 2023, Khanmigo is an AI-powered tutoring and teaching assistance initiative developed by Khan Academy, a nonprofit educational organization providing free learning resources worldwide. Khanmigo uses generative AI to support personalized instruction, student learning, teacher preparation, and classroom activities. Unlike general-purpose AI chatbots, Khanmigo is designed specifically for educational applications. It emphasizes guided learning, helping students work through problems, develop reasoning skills, and understand concepts rather than simply providing answers. The platform integrates AI capabilities with Khan Academy's instructional content, exercises, and student learning records.

Khanmigo has expanded from an experimental tutoring tool into an increasingly integrated component of Khan Academy's digital learning platform. Its development reflects growing interest in combining AI with established instructional practices, human teaching expertise, and educational research.

The name, Khanmigo, is derived from the Spanish word conmigo which means "with me" and the word Khan for Khan Academy.

Key Activities

  • Artificial Intelligence-Powered Student Tutoring
    • Personalized instructional assistance across subjects including mathematics, science, writing, history, and computer programming.
    • Through interactive conversations, it helps students understand concepts, work through problems, and develop independent reasoning and problem-solving skills. Its instructional approach emphasizes guided questioning and feedback, with support designed around students' learning needs.
  • AI Teaching Assistance
    • Provides educators with tools to develop standards-aligned lesson plans, instructional activities, learning objectives, assessments, grading rubrics, and student feedback.
    • Teachers can use the platform to develop differentiated instructional materials, generate practice questions, and identify areas where students may require additional assistance.
  • Integration with Classroom Learning
    • In 2026, introduced a redesigned learning experience that embeds Khanmigo more directly into student practice activities.
    • Rather than requiring students to open a separate AI tutoring application, the redesigned platform can provide assistance when students encounter difficulties or need additional explanation.
    • Changes were informed by pilots involving approximately 5,100 students across five U.S. school districts.
  • Interactive Visual Learning
    • Through a 2026 partnership with Google.org, Khan Academy introduced new capabilities using Google's Gemini AI models.
    • Khanmigo can generate interactive diagrams for mathematics and science instruction, allowing students to manipulate visual elements and receive feedback based on their actions.
    • These capabilities extend tutoring beyond text-based conversations.
  • Targeted Practice and Instructional Support
    • Khanmigo helps teachers generate customized practice questions that can be reviewed, modified, and assigned to students.
    • 2026 enhancements include stronger connections between AI-generated materials, teacher oversight, student practice, and instructional reporting.
  • Writing and Communication Support
    • Khanmigo includes tools that help learners develop writing skills through guided feedback, revision, and structured assistance.
    • Its Writing Coach functionality allows students to work through the writing process while providing educators with information about student progress.
  • School and District Implementation
    • Khan Academy works with school districts to integrate Khanmigo with instructional programs, classroom management processes, and learning analytics.
    • District implementations provide access to student tutoring, teacher assistance, administrative reporting, and professional development.
    • The initiative supports integration with existing educational technology environments, including the Canvas learning management system.
  • International Expansion
    • Khanmigo for Teachers has expanded beyond the U.S.
    • By December 2025, Khan Academy reported its teacher tools were available in more than 180 countries and territories, with expanded language support.
    • Availability and access arrangements differ for teachers, individual learners, families, and school districts.
  • Responsible AI Development
    • Khan Academy has developed a framework for responsible AI I education.
    • The approach addresses student safety, educational effectiveness, privacy, human oversight, and the risk of inaccurate or inappropriate AI-generated information.
    • Educators are encouraged to review generated materials and retain responsibility for instructional decisions.

Background

Khan Academy launched Khanmigo during a period of rapid development in generative AI. The initiative was designed to explore whether AI could make individualized tutoring more widely accessible while helping teachers manage instructional responsibilities. Early development emphasized the potential of conversational AI to provide personalized academic assistance. Classroom implementation revealed that access to AI did not necessarily lead students to use it frequently or effectively. In response, Khan Academy began redesigning the learning experience to integrate AI more directly into instructional activities.

The organization's 2025–2026 development work emphasized combining personalized tutoring with established content, student practice, learning analytics, and teacher oversight. By the beginning of the 2026–2027 school year, Khan Academy was rolling out a redesigned experience to district partners, including more proactive Khanmigo support, clearer learning pathways, and improved teacher and administrator dashboards.

Evaluation & Evidence

Khan Academy has undertaken research and product testing to evaluate Khanmigo's instructional effectiveness, student engagement, response accuracy, and performance. Between October 2025 and April 2026, Khan Academy conducted approximately 20 substantive product tests involving more than 15 million tutoring conversation threads. Studies examined how changes in tutoring design, access to student learning histories, and response processes influenced performance.

The organization reported that providing Khanmigo with structured information about students' recent practice and prerequisite learning improved performance on subsequent problems. Collectively, the tested improvements were associated with a 6.1% improvement in next-item correctness, a measure of whether students could correctly answer a subsequent problem independently after receiving tutoring.

Independent research has also examined Khanmigo in school settings. An August 2026 National Bureau of Economic Research working paper, One Click Away: AI Tutoring with Khanmigo in a Two-Year School Experiment, reported findings from a randomized study involving 18 Tennessee middle schools. The researchers identified modest improvements in math achievement among students assigned to an instructional program using Khan Academy with Khanmigo. However, students often made limited use of the AI tutor, suggesting that access alone may not be sufficient to produce meaningful learning benefits.

The study did not establish that Khanmigo independently caused the observed achievement improvements. Much of the measured benefit appeared related to the broader Khan Academy practice experience.

Together, these findings reinforce the importance of instructional design, student engagement, teacher involvement, and integration with learning activities. They also demonstrate the need to distinguish improvements in short-term task performance from independently verified long-term learning outcomes.

Khanmigo's development remains ongoing. AI-generated responses can contain errors, and student learning outcomes may vary according to implementation practices, educational setting, student characteristics, and the extent of engagement with the technology.

Partners & Funders

  • Khan Academy develops and operates Khanmigo as part of its nonprofit educational mission.
  • Google.org has supported Khanmigo development through philanthropic funding and technical expertise. A 2026 collaboration involving Google.org Fellows and Google's Gemini models contributed to interactive visual tutoring and targeted practice-generation features.
  • Microsoft has supported Khanmigo through technology infrastructure and collaboration, including efforts to make Khanmigo for Teachers available without charge to educators in eligible countries.
  • School Districts and Educational Institutions participate in Khanmigo implementation, testing, evaluation, and instructional development. These partnerships have informed changes in tutoring design, classroom integration, and educational reporting.
  • Instructure, provider of the Canvas learning management system, has collaborated with Khan Academy to make Khanmigo teacher tools accessible within the Canvas environment.
  • Khan Academy's broader philanthropic supporters contribute to the organization's educational mission, although not all are specifically identified as funders of Khanmigo.

Resources

Khan Academy. (n.d.). Khanmigo: Artificial intelligence-powered tutoring and teaching assistance. https://khanmigo.ai/

Khan Academy. (2024, May 22). A note from Sal Khan: Why we're offering Khanmigo for Teachers for free. https://blog.khanacademy.org/a-note-from-sal-khan-why-were-offering-khanmigo-for-teachers-for-free/

Khan Academy. (2026). How Khan Academy is building a better AI tutor: Our most recent learnings. https://blog.khanacademy.org/how-khan-academy-is-building-a-better-ai-tutor-our-most-recent-learnings/

Khan Academy. (2026, June 15). Built in the open: How pilot districts shaped the reimagined Khan Academy. https://blog.khanacademy.org/built-in-the-open-how-pilot-districts-shaped-the-reimagined-khan-academy/

Khan Academy. (2026, August 27). New AI tools bring interactive diagrams and targeted practice thanks to Khan Academy's partnership with Google.org. https://blog.khanacademy.org/new-ai-tools-bring-interactive-diagrams-and-targeted-practice-thanks-to-khan-academys-partnership-with-google-org/

Khan Academy. (n.d.). Khan Academy's framework for responsible AI in education. https://blog.khanacademy.org/khan-academys-framework-for-responsible-ai-in-education/

Oreopoulos, P., & Low, N. (2026). One click away: AI tutoring with Khanmigo in a two-year school experiment (Working Paper No. 35620). National Bureau of Economic Research. https://doi.org/10.3386/w35620

 

 

 

Relational Map: Key Connections within the Learn-and-Work Ecosystem

Research & Development: Khanmigo: Artificial Intelligence-Powered Tutoring and Teaching Assistance — Khan Academy

Initiative in the Learn-And-Work Ecosystem

The AI Library Mapping Assistant is a work in progress. Users are advised to verify the information from additional sources.
We strive for accuracy but acknowledge that errors may occur, and we encourage users to provide constructive feedback. 

Request an Edit

Have something to add or refine? Your input in this work matters greatly and we look forward to reviewing your additions

How useful was this resource?

Click on a star to rate it!

★ ★ ★ ★ ★
No votes so far! Be the first to rate this resource.

Organizations (561)

Initiatives (784)

Topic Briefs (163)