Learn Responsible AI: Verification, Privacy, and Human Review
Minimize the data you provide
Use the least information necessary for the task. Do not submit personal, confidential, health, financial, student, or client data unless there is clear authorization, an approved tool, and an understood retention and access policy. Fictional or redacted examples are suitable for practice.
Check output quality and fairness
Review claims against reliable sources. Look for omissions, stereotypes, uneven treatment, inaccessible language, and assumptions about people or groups. The tool's confident tone does not establish that a result is fair or correct.
Respect rights and disclose assistance when required
Confirm permission for source materials, images, names, and other protected content. Follow the current service terms, organization policy, and context-specific disclosure requirements. Do not present a generated quotation, testimonial, or result as if it came from a real person.
Keep a human accountable
Assign a reviewer who can understand the task and correct the output. Record material edits and limitations. Stop and escalate when the tool produces sensitive information, a high-impact recommendation, or an answer that cannot be verified.