GenZAgents for edtech — student-data-safe AI audit + accessibility evidence

FERPA, UK GDPR, COPPA each restrict how student data flows to third-party AI providers. Receipts capture which AI provider processed which class of data — your audit trail proves COPPA-compliant providers were used for under-13 data; UK GDPR DPA mapping per provider; FERPA "directory information" boundaries respected. Without receipts: hope; with: evidence.

AI systems for "determining access to educational institutions" or "evaluating learning outcomes" are classified high-risk under EU AI Act Annex III. High-risk systems have extensive obligations (Articles 9-15): risk management, data governance, technical documentation, record-keeping, transparency. The record-keeping requirement (Article 12) explicitly calls for AI activity logs — receipts satisfy this natively.

AI grading / placement / admissions assistance must be monitored for bias against protected groups. The receipt feed lets you query AI decisions by demographic dimensions (if your edtech captures those). The dashboard's bias-monitoring views (Enterprise tier) flag distributions that diverge between demographic groups — an early-warning signal before the regulator gets there.

AI-generated educational content must satisfy WCAG / EN 301 549. The receipt captures the model used + the accessibility-pass flag (set if the post-AI content was validated against WCAG). Audit trail for accessibility compliance becomes a single dashboard view.

AI tutor interacts with students. Each tutoring session is a receipt — student ID (with consent flow audited separately), session topic, AI model, summary of interactions. The receipt feed lets parents / teachers / district admins query their authorised slice of student AI activity. FERPA-compliant access controls govern who sees what.

AI-graded essay rubric application. Receipt captures: student ID (de-identified for AI provider), essay digest, AI grade, supervising teacher human_id, teacher's final grade. Discrepancies between AI grade and teacher grade are flagged — useful both for quality control (catching AI errors) and for understanding teacher-AI interaction patterns.