Responsible AI Checklist — aiborg™ Resource Library

Completion Progress0 / 22 items

⚖️ 1. Fairness

Bias testing completed — Model outputs tested across demographic groups for disparate impact Training data audited — Data sources reviewed for representational balance and historical bias Fairness metrics defined — Quantitative thresholds set for demographic parity or equalised odds Mitigation plan documented — Steps to address identified biases before and after deployment

🔍 2. Transparency

Model card published — Intended use, limitations, and performance benchmarks documented Decision explainability — Users can understand why the AI made a particular recommendation AI disclosure — Users are clearly informed when they are interacting with an AI system

📋 3. Accountability

Ownership assigned — A named individual is accountable for the AI system's outcomes Audit trail maintained — Decisions and model changes are logged for retrospective review Incident response plan — Process defined for handling AI failures, errors, or harmful outputs Regular review schedule — Recurring audits are calendared (quarterly or after major changes)

🔒 4. Privacy

Data minimisation applied — Only data necessary for the task is collected and processed PII handling reviewed — Personal data anonymised, encrypted, or excluded as appropriate Consent mechanisms — Users informed about data usage and given meaningful consent choices Regulatory compliance — GDPR, DPDP Act, or relevant data protection law requirements met

🛡️ 5. Safety

Red-teaming conducted — Adversarial testing performed to identify harmful failure modes Output guardrails set — Content filters, rate limits, and output validation in place Fallback mechanisms — Graceful degradation path when the AI cannot provide a reliable answer

👤 6. Human Oversight

Human-in-the-loop defined — Critical decisions require human review before action is taken Override capability — Operators can override, pause, or shut down the AI system at any time Escalation path clear — Users know how to escalate concerns about AI behaviour

Pre-Deployment Sign-Off

Full Name

Role / Title

Date

Signature