Atlas for School: AI Cheating Prevention Framework
1. Guardian Mode: Classroom Integrity AI
Real-time monitoring system active during assessments.
Flags behavior patterns, external device usage, or suspicious browser activity.
Uses zero-retention protocols—no permanent data storage unless cheating is detected.
2. Sovereign ID Verification
Students must timestamp their ID (photo or keystroke biometric) to verify identity before any major exam.
Uses Atlas Layer to store proof-of-authorship without compromising privacy.
Prevents impersonation or AI-generated submission fraud.
3. Proof of Learning Protocol (PLP)
Every assignment must pass through Atlas PLP, which scans for:
AI-generated content patterns.
Unnatural structure shifts or writing inconsistency.
Non-human timestamp gaps between input and submission.
If flagged, the student is prompted to defend or revise their work with transparency.
4. Covenant Honor Agreement
Before any assessment, students agree to a Digital Honor Covenant.
The covenant is timestamped on the Atlas Layer and binds them to a code of ethics.
Violating this code results in:
Suspension from AI tools.
Triggered parental and instructor notification.
Optional restorative pathway instead of punitive measures.
5. Offline Mode + Tamper Shield
Tests or essays can be run in Atlas Shield Mode:
No AI access allowed (local lockdown browser or device setting).
AI-generated tools like auto-complete and autocorrect disabled.
Any bypass attempt triggers a silent Guardian Alert.
6. Educator Console
Teachers get a real-time dashboard with:
Cheating probability scores.
Behavior logs during the test.
Timestamp trail of student input patterns.
Uses AI to assist—not replace—human judgment.
7. Redeemable Pathways (Optional but Encouraged)
If a student cheats, Atlas allows one Redemption Assignment:
A guided project on ethics, responsibility, and how AI works.
Successful completion reinstates access.