Program Structure

Module 1: Ethics in AI (What Matters)

• Common harms: discrimination, surveillance, misinformation, unsafe automation.
• Stakeholders: users, affected groups, operators, regulators.
• High-risk vs low-risk AI and when to avoid automation.
• Ethics workflow: identify risk → controls → evidence → review.

Module 2: Policy Foundations

• Policy vs standards vs SOPs vs guidelines.
• Define scope: systems covered, data types, model types, vendors.
• Roles: owner, approver, reviewer, legal, security, product.
• Decision rules: approvals, exceptions, escalation.

Module 3: Data Governance, Privacy, and Consent

• Data mapping: sources, sensitivity, access, retention.
• Privacy controls: minimization, purpose limits, anonymization (intro).
• Consent and notice basics; handling sensitive data.
• Security basics: access, logging, incident response (overview).

Module 4: Fairness, Bias, and Human Impact

• Bias types: data, measurement, selection, labeling.
• Fairness checks: group metrics, error breakdowns.
• Human-in-the-loop: when required and how to design it.
• Impact assessment: who is affected and how.

Module 5: Transparency and Explainability

• What to disclose: purpose, limits, confidence, sources.
• Explainability levels: simple models vs black-box models (overview).
• Documentation: model cards, datasheets, change logs.
• Communication: user-facing AI notices and disclaimers.

Module 6: Safety, Misuse, and Security

• Threats: prompt injection, data leakage, jailbreaks (intro), model abuse.
• Content risks: hallucinations, toxic outputs, sensitive info exposure.
• Controls: access limits, filtering, rate limits, review workflows.
• Red-teaming basics and incident playbooks.

Module 7: Compliance and Audit Readiness

• Compliance mindset: evidence, traceability, accountability.
• Lifecycle gates: design → build → test → deploy → monitor.
• Monitoring: drift, performance, complaints, harm signals.
• Audit pack: approvals, tests, risk register, logs.

Module 8: Policy Pack Building (Templates)

• AI Use Policy: allowed uses, restricted uses, prohibited uses.
• Risk Assessment: severity/likelihood, mitigations, owners.
• Model Governance: testing checklist, release checklist, monitoring plan.
• Vendor Policy (intro): due diligence questions and contract clauses (overview).

Final Project

• Choose a use case: HR screening, customer support bot, fraud, healthcare analytics (non-clinical), education.
• Deliverables: AI policy + risk assessment + governance checklist + transparency notice.
• Optional: short presentation for leadership review.