AI Data Privacy Statement

Lessi AI is dedicated to responsibly developing and deploying AI-powered educational tools that respect user privacy, promote ethical outcomes, and remain transparent and accountable. By leveraging Microsoft Azure AI services and aligning with Microsoft’s Responsible AI Principles, we ensure fairness, reliability, security, inclusiveness, transparency, and accountability, all while fully complying with the Family Educational Rights and Privacy Act (FERPA).

Foundational Models

We host pre‑trained foundation models in Azure AI Foundry. These models are managed by Microsoft within Azure’s secure environment but may originate from various third‑party providers (OpenAI, Hugging Face, Meta, etc.).

No Student Data Used for Training

Lessi AI never uses student or teacher interaction data to train, fine‑tune, or improve any large language model. Prompts and responses are processed in real time and are not retained or repurposed for model training.

Data Isolation & Ownership

All data remains within Lessi's Azure environment. Microsoft’s Azure OpenAI Service does not share your content with external model providers; all data remains within your control and is processed in the United States.

• Azure OpenAI Data Privacy Policy
• Content Filter
• Default Safety Policies

Encryption & Security Controls

Lessi encrypts all data both in transit (TLS/SSL) and at rest (AES‑256). Keys are managed in Azure Key Vault, and network isolation is enforced via Azure Private Link. Lessi AI’s access is governed by Azure Active Directory and role‑based access control (RBAC).

Auditability & Traceability

Azure Monitor and built‑in audit logs capture every access and modification event. Detailed traceability ensures accountability in data processing and supports FERPA’s right to review and amend records.

We follow Microsoft’s Responsible AI Standard to guide our design, development, and deployment practices.

Fairness

• We use Azure AI’s fairness assessment tools to detect and mitigate bias in training data and AI outputs.
• Datasets are continually refined to represent diverse student populations and learning styles.
• Educators and domain experts validate AI‑generated recommendations to ensure equitable treatment.

Reliability & Safety

• Automated testing, continuous monitoring, and validation processes ensure high availability and correct behavior.
• Human oversight is applied to high‑impact decisions, giving educators final control.
• Threat detection and cybersecurity protocols protect against malicious actions and unintended AI behaviors.

Privacy & Security

• Data Minimization: We collect only essential data and anonymize PII wherever possible.
• Encryption: All data is encrypted in transit and at rest; keys stored in Azure Key Vault.
• Access Controls: Strict RBAC policies limit data access to authorized personnel.
• Compliance: Practices align with GDPR, FERPA, and other applicable laws.
• Incident Response: We maintain a formal plan to detect, investigate, and remediate security incidents.

Inclusiveness

• AI‑powered accessibility features support users with impairments.
• Compliance with WCAG 2.1 ensures an equitable digital learning experience.
• Collaboration with educators tailors recommendations for neurodiverse and special‑needs learners.

Transparency

• We provide clear explanations of how AI recommendations are generated.
• Decision‑making processes and limitations are documented and accessible to stakeholders.
• Audit logs and reports are available for review by educators and administrators.

Accountability

• Routine audits verify adherence to responsible AI and FERPA requirements.
• User feedback channels allow reporting of AI concerns or unexpected behaviors.
• Escalation procedures ensure timely investigation and resolution of issues.
• All staff receive training on AI ethics, bias mitigation, and data privacy protocols.

Lessi AI integrates Microsoft Responsible AI Principles with rigorous FERPA compliance to deliver a safe, fair, and transparent AI platform for educators. We never use student data for model training, all operations occur within Azure’s secure environment, and robust controls ensure privacy, security, and accountability. For more information, please contact our compliance team or review Microsoft’s documentation on Azure AI and data privacy.