SafeGradeAI
The independent AI certification authority

Become A Member of the SafeGradeAI Movement

We are the independent authority helping schools, parents, and vendors adopt AI responsibly, transparently, and safely. Here's who we are, the standards behind every grade we publish, and how to join.

Independent · No vendor pay-for-pass Methodology v2.5 · publicly versioned Continuously re-evaluated

Who we are

Protect students. Empower educators. Hold AI accountable.

SafeGradeAI exists because AI tools are reaching children faster than schools can vet them. We provide the independent evaluations, certification, and policy frameworks that make safer adoption possible — without slowing the pace of innovation.

Our mission

A trusted line between AI and the people responsible for kids.

Independent reviews, classroom-ready guidance, and certifications that schools and families can actually act on.

Our vision

The standard for AI in K-12.

The authority schools, vendors, and families rely on to know which AI tools have earned classroom trust — and which haven't yet.

What sets us apart

Independence is the product

Independent evaluations

We are not paid by the vendors we evaluate. Period. Conflicts of interest are disclosed publicly.

Education-focused standards

Our framework is built for classrooms — not generic adult productivity benchmarks.

Parent-and-school centered

Every report includes both a school implementation note and a plain-English parent summary.

Transparent results

Every score links to the evidence behind it, with clear reasoning you can audit and defend.

Compliance-grade citations

COPPA, FERPA, GDPR-K, and UK AADC references at the subcategory level — not vague hand-waving.

Practical implementation

Every report ends with what to actually do in classrooms next week — not a 50-page PDF nobody reads.

The Framework

Five evaluation dimensions, weighted for educational use

SafeGradeAI uses a proprietary multi-factor evaluation framework. Each certified product is reviewed against five public dimensions, weighted by impact on student outcomes. Specific sub-criteria and weights are intentionally not disclosed to preserve framework integrity.

Privacy & Data Handling

Moderation & Safety Controls

Age Appropriateness

Transparency & Governance

Educational Alignment

A live score, not a one-time stamp

Certifications are versioned, dated, and re-scored on a defined cadence. When a vendor's safety posture changes, we publish a downgrade alert — even when it's inconvenient for the vendor.

92/100
Sample SafeGrade
KidSafe

Meet the platform

How a SafeGradeAI evaluation gets made

1

Research

Reviewers study the tool's documentation, privacy policies, and real-world behavior.

2

Score

Each subcategory is scored using our proprietary framework with cited evidence.

3

Peer review

Two independent reviewers verify findings before any score is finalized.

4

Publish & monitor

Reports go live with a version number and stay under continuous monitoring.

Trust Center

Our nine commitments

The principles that govern every evaluation, every published grade, and every alert we send.

01

Our Commitment to AI Safety

The safety of children using AI is not a feature of our product — it is the entire reason it exists. Every grade we publish, every certification we issue, and every alert we send is measured against this single standard.

02

Evaluation Integrity

We do not negotiate scores. Vendors cannot review, appeal, or pay to change a published grade. Submission fees cover evaluation cost and never influence the verdict.

03

Transparency Standards

Our methodology is publicly versioned (v2.5). Every certification page cites framework dimensions, evidence sources, and last-evaluated date. Where we redact internal criteria, we say so explicitly.

04

Educational Responsibility

Schools and districts adopt AI tools that shape young minds. We publish guidance that is classroom-ready, board-defensible, and aligned to how educators actually make decisions.

05

Parent Trust Principles

Parents are the last line of defense in a child's AI exposure. Reports are written in plain language, include conversation scripts and configuration walkthroughs, and never require a CS degree to act on.

06

School Trust Principles

Educators and administrators need legally-defensible decisions. Every school-tier evaluation includes COPPA, FERPA, GDPR-K and UK AADC references — with cited sections, not blanket claims.

07

Data & Privacy Commitments

We collect only what we need to operate. We do not sell visitor or member data. We do not run third-party ad networks on logged-in experiences. Payment data is handled by Stripe.

08

Independent Evaluation Standards

No vendor is granted preferential treatment. Sponsored evaluations are clearly labeled, follow the identical published rubric, and the published grade is final. We will fail a sponsoring vendor.

09

Continuous Review Process

An AI tool's safety posture in January may not hold in March. Certified tools are monitored 24/7 — when policies, models, or behaviors change, we re-evaluate and publish a real-time alert to Premium members.

Security & Privacy

How we handle your data

The infrastructure, posture, and constraints behind the platform.

Encryption in transit & at rest

All traffic is served over TLS. Data is stored with encryption at rest. Backups are encrypted.

Least-privilege access

Production data access is limited to a small set of admins, audited, and protected by SSO + 2FA.

Row-level security

Every user-facing table uses Postgres row-level security so users only ever see their own records.

No selling of personal data

We do not sell visitor or member data. We do not run third-party ad networks on logged-in experiences.

COPPA-aware design

SafeGradeAI is not directed to children under 13. Member accounts are for adults (parents, educators, administrators).

FERPA-aware reviews

When evaluating school-facing tools, we surface each vendor's posture on FERPA and major state student-privacy laws.

Reputable subprocessors

We use established providers for hosting, database, payments (Stripe), email, and analytics.

Independent of vendors

We do not let vendors edit their own scores. Sponsored evaluations are clearly labeled and follow the published rubric.

What we collect

  • Account data: Email, name, account type (parent / school / vendor), organization name.
  • Usage data: Aggregate analytics on which evaluations and pages are visited. Never sold.
  • Submissions: Tool review requests, sales leads, contact messages, and tool submissions you choose to send us.
  • Payment data: Handled by Stripe. We never see or store full card numbers.

COPPA & FERPA posture

  • COPPA: SafeGradeAI is not directed to children under 13. Member accounts are for adults (parents, educators, administrators). We do not knowingly collect personal information from children under 13.
  • FERPA: When evaluating school-facing tools, we surface each vendor's FERPA posture. We do not process student education records ourselves.
  • Reviews, not a service to kids: We publish safety evaluations and guidance. We are not an ed-tech product used directly by students in classrooms.

Reporting a security concern

If you believe you have found a security issue, please email support@safegradeai.com. We respond to legitimate reports within two business days.

Premium membership

Become A Member — $9/month

Premium is one membership for parents, schools, and vendors. Unlock the full Download Center, the live Safety Updates feed, unlimited Watchlists, the Personal AI Safety Advisor, and every printable playbook our editorial team publishes.

Full Download Center library
Real-time Safety Updates feed
Unlimited tool Watchlists
Personal AI Safety Advisor
Per-tool Safety Brief PDFs
Re-evaluation & downgrade alerts

Work with us

Transparency runs both ways

Have an AI tool to submit, a partnership in mind, or research to contribute? Reach out.

Request an evaluation

Vendors and schools can submit AI tools for SafeGradeAI review.

Submit a tool

Partnership opportunities

Districts, EdTech organizations, and research labs interested in co-publishing.

Contact Us

Research contributions

Academic researchers and safety teams welcome to contribute to our evaluation framework.

Read our blog

By subscribing you agree to receive emails from SafeGradeAI. You can unsubscribe at any time.