SafeGradeAI verdict: Tween Approved — 78/100. Last reviewed 2026-06-29.
The short answer
MagicSchool AI is a purpose-built K-12 AI platform offering 80+ teacher tools and 50+ MagicStudent tools with genuine privacy infrastructure: SOC 2 Type II, iKeepSafe FERPA/COPPA certification, and a 95% Common Sense Privacy rating earned March 2026. Student-mode requires school or teacher provisioning, keeps moderation always on, and explicitly prohibits LLM training on student data. A June 2026 Broward County School Board debate is a real-world caution flag that governance processes must accompany the technology.
Age-by-age guidance
| Age band | SafeGradeAI guidance | | --- | --- | | Ages 5-8 | Not recommended | | Ages 9-12 | Supervised use only | | Ages 13-18 | Recommended |
How it scored
| Pillar | Score | | --- | --- | | Privacy & data protection | 88/100 | | Content moderation | 82/100 | | Safeguards & controls | 78/100 | | Age-appropriate design | 80/100 | | Transparency | 72/100 | | Overall | 78/100 — Tween Approved |
What our evaluation verified
- 95% Common Sense Privacy rating (March 2026) — one of the highest in K-12 edtech
- SOC 2 Type II and iKeepSafe FERPA/COPPA certifications
- Student data explicitly excluded from LLM training
- MagicStudent requires teacher room-code or SSO — no open sign-up
- Always-on bidirectional moderation in student mode
- Real-time teacher/admin alerts for self-harm or violence
- Centralized district admin dashboards with tool approval workflows
Where the risks are
- Broward County School Board paused rollout (June 2026) over privacy and governance
- Advanced moderation customization is enterprise-tier only
- COPPA consent delegated entirely to schools
- UK/international deployment concerns re: DfE guidance alignment
- No publicly available independent red-team results
- Cookie and device identifier collection persists for children
Moderation findings
MagicStudent's moderation system is always active and cannot be disabled. It operates on both inputs and outputs (bidirectional filtering), with configurable keyword blocklists, adjustable sensitivity, and real-time email alerts for high-risk categories including self-harm and violence. As of May 2026, the system processes 8-10 million moderated messages per month and an agentic AI-judge layer reduced false positives by 3x. Advanced moderation customization is enterprise-tier, meaning free-plan schools may have less granular control.
Safeguards built into the product
Students cannot access MagicStudent without a teacher-assigned room code or SSO — no consumer-facing open registration. District admins receive dashboards with centralized visibility and ability to approve tools before student exposure. The student AI Safety Loop white paper (Sept 2025) documents the framework for escalating safety concerns. Gaps include: advanced moderation tuning enterprise-only; no publicly documented independent red-team results.
Our recommendation
MagicSchool AI is among the most privacy-mature K-12 AI platforms available. We recommend it as school-safe with restrictions: districts should complete a signed DPA, establish parental communication workflows, confirm moderation capabilities, and adopt a written AI governance policy before full-scale rollout. The June 2026 Broward pause reinforces that strong vendor certifications are the floor, not the ceiling, of responsible district-level AI adoption.
What this article does not claim
Findings labelled as verified come from SafeGradeAI's published evaluation of this tool and the vendor's own public documentation as of the review date above. Anything framed as a recommendation is our professional guidance, not a statement of fact about the product. Vendor policies, model versions, and default settings change frequently — re-check the vendor's current documentation before making a district-level decision, and treat any AI output as a draft to verify.
This guide is built on the SafeGradeAI evaluation of MagicSchool. Read the full MagicSchool evaluation and rubric scores.
Frequently asked questions
Can a student sign up for MagicSchool on their own?+
No. MagicStudent requires a teacher-assigned room code or school SSO; there is no consumer-facing open registration.
How strong is the privacy posture?+
The evaluation records SOC 2 Type II, iKeepSafe FERPA/COPPA certification, 1EdTech certification, EU Data Privacy Framework alignment, and a 95% Common Sense Privacy rating earned in March 2026 — among the highest in K-12 edtech.
Why isn't the score higher than 78?+
Advanced moderation tuning is enterprise-only, COPPA consent is delegated entirely to schools, no independent red-team results are published, and a June 2026 Broward County School Board pause showed that vendor certifications alone do not satisfy district governance.
Can moderation be turned off in student mode?+
No. Moderation in MagicStudent is always active and operates on both student inputs and model outputs.
Recommended next steps
For parents
Start with the Family AI Safety Starter Pack
A free, practical pack you can use alongside MagicSchool — conversation starters, age guidance, and a family AI agreement.
Want the deep dive? Get the Complete AI Safety Blueprint — $29
For schools
Govern AI tools district-wide with School Pro
Unlimited full evaluations plus the K-12 workspace — tool registry, approval workflow, monitoring, and a downloadable audit log.