Operational guidance for human-in-the-loop design, monitoring, incident response, privacy/compliance checkpoints, and lightweight model governance.
An actionable, interactive checklist to help teams detect, monitor, and mitigate fairness, bias, and ethics risks in models and data pipelines. Use this form to record current monitoring status, select metrics, document mitigation steps, assign ownership, and preserve an audit-friendly record.
Operational runbook to detect, triage, investigate, communicate, contain, remediate, and learn from incidents caused by AI outputs, model failures, or AI-enabled pipelines. Includes role assignments, triage checklist, evidence-gathering template, escalation paths, communication templates, immediate containment and rollback steps, and a post-incident remediation and monitoring plan.
Plug-and-play workflows, ready prompts, safety guardrails, and a short supervisor checklist frontline teams can use to save time, reduce errors, and keep human judgment where it matters.
Actionable runbook for detecting, triaging, containing, and investigating incidents that involve AI models or AI outputs. Includes classification guidance, step-by-step triage and evidence collection checklists, containment and rollback options, communication templates, forensic indicators for common AI incident types, monitoring suggestions, roles and responsibilities, and post-incident remediation steps.
An actionable, interactive checklist to guide data minimization, pseudonymization, safe sampling, retention scheduling, and privacy review sign-off for AI projects. Designed for project teams to record decisions, attach artifacts, and produce a stored sign-off record.