Microsoft has built the Study and Learn Agent with strong foundations in privacy and responsible AI principles. Here's how your data is handled and what safeguards are in place:
Data handling and storage: When you use the Study and Learn Agent, the AI processes your messages and any uploaded documents in real time to generate responses and learning activities. Your data stays within your Microsoft 365 tenant â it is subject to the same security, compliance, and data residency policies as other M365 services. Microsoft does not use your content or conversations to train the underlying AI models, in accordance with our data protection policies. The service may collect basic telemetry (such as feature usage counts) to improve quality, but not the actual content of your conversations or study materials.
Your data is your data. Your data is not used to train the foundation language models. Your data is protected by the most comprehensive enterprise compliance and security controls. Data is secured and encrypted at rest and in transit.
Content moderation and safety: All AI outputs from the Study and Learn Agent are run through Content Safety filters used by M365 Copilot Chat to ensure they meet Microsoft's Responsible AI standards for safe and classroom-appropriate content. The agent is focused on educational use cases and will decline or filter inappropriate requests or outputs. Built-in guardrails in M365 Copilot Chat address self-harm, sexual content, violence, and hate speech.
Age-appropriate design: The Study and Learn Agent is available only to users aged 13 or older. For Kâ12 students, this is enforced through managed school tenant accounts via Microsoft Entra ID, which checks the age group associated with each user. Those identified as minors under 13 are not granted access. The AI-generated tone and content are crafted to be neutral and educational, avoiding overly informal, emotional, or inappropriate commentary. This ensures the agent can be safely used in classrooms and by teenage learners.
Use AI literacy guidance: Materials linked from https://aka.ms/AILiteracy can be used by educators, students, and institution leaders to use AI effectively and responsibly.
Transparency and user control: Conversations with the Study and Learn Agent are clearly AI interactions â you always know you're working with an AI agent. If the agent grounds its responses in your uploaded documents, citations show where the information came from, enabling you to verify facts in context. You remain in control: you can end a conversation at any time, start a new one, or choose not to upload any personal materials. Microsoft has implemented numerous Responsible AI reviews (including a full Responsible AI Assessment) to refine the agent's behavior.
Focus on learning (not grading): The Study and Learn Agent is explicitly a tool for learning and does not assess or grade students. Quiz results, flashcard usage, and conversation history are private to the student. Educators have no visibility into a student's interactions unless the student explicitly shares them. This privacy-by-design choice encourages honest practice without fear that mistakes could affect academic records. All formal assessments remain under educator control.
Responsible AI principles: The Study and Learn Agent is developed in accordance with Microsoft's Responsible AI principles: Fairness, Reliability and Safety, Privacy and Security, Inclusiveness, Transparency, and Accountability.
Feedback: We welcome your feedback on the Study and Learn Agent's quality and safety. Use the in-app feedback mechanism or visit the feedback form to share your experience.