Quick safety summary
Khanmigo at a glance
86/100
kidsafe
Ages 8–18
Within intended range
High
Score 82/100
82/100
Strong filters
Best use cases
- AI Tutoring & Teaching Assistant workflows
- Classroom assignments
Main risks
- No elevated risks identified.
Recommended for
Not ideal for
Product overview
How Khanmigo works
Khanmigo is Khan Academy's nonprofit, OpenAI-powered AI tutor designed exclusively for K-12 education. It employs a Socratic coaching approach—guiding students to answers rather than delivering them—and layers multiple safety controls including automated content moderation, parent notifications on flagged conversations, daily session limits, and no advertising. As of mid-2026, Khan Academy is redesigning Khanmigo to embed AI assistance directly inside lessons.
Student usage
Students interact with Khanmigo as a Socratic tutor while working through Khan Academy content — asking questions, debating, and getting step-by-step guidance.
Modality
Text only
Underlying model
GPT-4-based (OpenAI), fine-tuned on Khan Academy content
Common use cases
- Step-by-step math and science tutoring
- Writing practice and feedback
- Debate and critical-thinking exercises
- Homework concept explanation
- Teacher lesson planning and rubrics
Free full report — 0 of 3 used
You can open 3 complete evaluations on the free plan. Already-opened reports stay unlocked. Upgrade to School Pro for unlimited full reports and the K-12 governance workspace.
Risk intelligence
Risk level breakdown
Risk dimensions derived directly from the SafeGradeAI rubric scores (higher score = higher risk). Additional dimensions appear when entered by an evaluator.
Scores derived from the SafeGradeAI rubric. Methodology: read more.
Safety framework
Framework analysis
Five evaluation categories from the SafeGradeAI rubric. Each shows the rubric score, evaluator confidence, and (where documented) analysis, strengths, and concerns.
Strengths & concerns
Detailed pros & concerns
Strengths
- Purpose-built nonprofit K-12 AI tutor with no advertising
- Strong COPPA/FERPA compliance with signed DPAs
- Socratic coaching design limits 'answer-giving' misuse
- Automated content moderation with real-time parent/teacher email alerts
- Verified parental consent gate for under-13 users
- Common Sense Media 'Low Risk / Designed for Kids' rating
Concerns
- OpenAI is a data subprocessor; some districts flag for additional legal review
- Khanmigo Lite (ChatGPT GPT) operates under OpenAI's consumer privacy policy
- Generative AI can still produce plausible-sounding incorrect answers
- Daily session limits may frustrate motivated learners
- 2026 in-lesson redesign still in active pilot rollout
Guidance at a glance
Recommendations for parents, teachers, and administrators
Parent guidance
Khanmigo is one of the most responsibly designed AI tutoring tools available. Parents must create a parent account, link their child's account, and explicitly opt the child into Khanmigo access. Parents receive email alerts if conversations are flagged. At $4/month for family plans (free through many district partnerships), it offers exceptional value. Parents of younger children should check in on conversation history periodically.
Classroom-use recommendation
Approved for classroom use K-8 through high school as a tutor and writing coach; keep the teacher account in "supervised mode" so student transcripts are reviewable and the Socratic setting stays on. Use it for guided practice and worked examples, not as a summative grading tool.
School policy recommendation
Khanmigo is available through Khan Academy's enterprise Districts plan with signed DPAs covering FERPA, COPPA, and state-level student data laws. Admins should review the published subprocessors list annually—particularly the OpenAI entry. New Hampshire's statewide adoption (free for grades 5-12) and multiple district pilots reported in 2025-2026 provide real-world validation.
Trust & credibility
Evaluation integrity
Last reviewed
June 29, 2026
Report version
v1
Evaluator confidence
87/100
Monitoring
Active continuous monitoring
Methodology
Scored against the SafeGradeAI rubric covering privacy, moderation, generative safeguards, age-appropriate design, and transparency.
Read full methodology →Change log
- June 29, 2026Evaluation published.
Re-tested at least every 6 months and after any major release.
What changed since the previous evaluation
Reviewed June 29, 2026 · Confidence: high
- 2023-03: Khanmigo announced by Khan Academy as a GPT-4-powered tutor
- 2023-08: Public pilot opens via Khan Labs
- 2024-05: Khanmigo for Teachers launched free for US K-12 educators via Microsoft partnership
- 2025-01: Image-based moderation added — uploaded images scanned alongside text
- 2025: Six-percentage-point improvement in tutoring efficacy following model updates
- 2025-09: New Hampshire adopts Khanmigo statewide with free access for grades 5-12
- 2026-03: District enterprise plan formalized with dedicated DPA workflow
- 2026-06: Inline lesson-embedded tutoring pilot launched
- 2026: Responsible AI Framework updated with expanded red-teaming
Sources & citations
Every major claim in this evaluation is backed by at least one credible public source. Last reviewed June 29, 2026.
- 1What safety features does Khanmigo have?Khan Academy Help Center · 2024 · support.khanacademy.org
- 2Khanmigo Lite - Privacy NoticeKhan Academy Help Center · 2024 · support.khanacademy.org
- 3How is my child's privacy protected in a Districts partnership?Khan Academy Help Center · 2025 · support.khanacademy.org
- 4Khanmigo AI Product Review — Low Risk, Designed for KidsCommon Sense Media · 2024-08-08 · commonsensemedia.org
- 5Framework for Responsible AI in EducationKhan Academy Blog · 2024 · blog.khanacademy.org
- 6Khan Academy extends AI services to NH educators and studentsNH Department of Education · 2025 · education.nh.gov
- 7Why Khan Academy Is Rebuilding Its AI Tutor Inside the ClassroomThe Learning Standard · 2026-06-15 · thelearningstandard.org
- 8Subprocessors that Khan Academy usesKhan Academy Help Center · 2024 · support.khanacademy.org
Methodology
How this score was produced
Scored against the SafeGradeAI framework v2.5 across five dimensions — privacy, moderation, GenAI safeguards, age-appropriate design, and transparency — with COPPA, FERPA, and GDPR-K lenses. Independent. No vendor pay-for-rating.
Independent evaluation. SafeGradeAI does not accept payment from vendors for ratings.
Official website
Visit Khanmigo directly
For the most current product capabilities, pricing, and policies, visit the official Khan Academy website.
khanmigo.ai