Ethics and Governance in the Age of Generative AI Course

Lesson 1 • Major Model Architectures and Modalities

Surveys transformer-based language models, diffusion image models, and multimodal systems. Enables learners to match governance strategies to specific technical architectures.

Lesson 2 • Capabilities, Limitations, and Failure Modes

Examines hallucination, context window constraints, and distributional bias as structural properties. Prepares learners to identify where ethical harms are most likely to emerge.

Lesson 3 • How Generative AI Produces Outputs

Covers probabilistic text and image generation, training data pipelines, and model inference. Grounds all subsequent ethics analysis in technical reality rather than speculation.

Lesson 1 • Navigating Ethical Conflicts and Trade-offs

Provides structured methods for resolving conflicts between competing ethical principles in real decisions. Prepares learners to justify choices transparently when values collide.

Lesson 2 • Classical Ethical Theories and AI

Introduces consequentialism, deontology, and virtue ethics as analytical lenses for AI harms and benefits. Establishes the theoretical base for all subsequent ethical reasoning in the course.

Lesson 3 • Stakeholder and Rights-Based Approaches

Analyzes how affected communities, marginalized groups, and future generations hold legitimate claims on AI systems. Grounds governance in human rights and participatory ethics.

Lesson 1 • Fairness Definitions and Their Tensions

Compares statistical fairness metrics—demographic parity, equalized odds, and calibration—and proves they cannot all be satisfied simultaneously. Enables informed selection of context-appropriate criteria.

Lesson 2 • Sources of Bias in AI Systems

Traces bias from data collection through labeling, model training, and deployment feedback loops. Establishes that bias is structural, not incidental, requiring systemic responses.

Lesson 3 • Mitigation Strategies Across the Pipeline

Presents pre-processing, in-processing, and post-processing interventions with their respective trade-offs. Connects mitigation choices back to the fairness criteria selected in prior sections.

Lesson 1 • Privacy Risks Unique to Generative AI

Covers memorization attacks, membership inference, and synthetic data re-identification. Connects these technical risks to concrete harms for individuals and organizations.

Lesson 2 • Privacy Concepts in the AI Context

Defines contextual integrity, informational self-determination, and the right to be forgotten as applied to AI training and inference. Establishes the normative baseline for data governance decisions.

Lesson 3 • Data Collection, Consent, and Provenance

Examines lawful bases for training data use, consent mechanisms, and data provenance tracking. Prepares learners to evaluate whether a dataset meets ethical and regulatory standards.

Lesson 1 • Explainability Methods for Generative AI

Surveys attribution methods, saliency maps, and chain-of-thought prompting as explainability tools. Addresses the unique challenge of explaining outputs from large generative models.

Lesson 2 • Transparency as a Governance Requirement

Distinguishes algorithmic transparency, process transparency, and disclosure obligations for different stakeholder audiences. Frames transparency as a precondition for meaningful accountability.

Lesson 3 • Accountability Structures and Responsibility Gaps

Analyzes how distributed AI supply chains create responsibility gaps and diffuse moral agency. Introduces structured accountability frameworks to assign roles and obligations clearly.

Lesson 1 • Content Safety and Output Controls

Examines classifiers, constitutional AI, and reinforcement learning from human feedback as safety mechanisms. Evaluates trade-offs between safety constraints and model utility.

Lesson 2 • Taxonomy of AI Harms

Classifies harms by type—physical, psychological, financial, societal—and by proximity of causation. Provides a shared vocabulary for risk assessment and policy design.

Lesson 3 • Risk Assessment Frameworks for AI

Applies probability-severity matrices, threat modeling, and red-teaming to generative AI deployments. Enables learners to produce structured risk registers for AI products.

Lesson 1 • Compliance Programs and Policy Development

Builds AI-specific compliance programs including policy hierarchies, training requirements, and monitoring cadences. Enables learners to draft enforceable internal AI use policies.

Lesson 2 • Organizational AI Governance Structures

Designs board-level oversight, AI ethics committees, and cross-functional review processes for AI deployment. Connects governance structures to accountability frameworks established in earlier chapters.

Lesson 3 • Voluntary Standards and Industry Frameworks

Examines international AI standards bodies, industry codes of conduct, and voluntary commitment frameworks. Helps learners leverage voluntary standards to build trust and anticipate future mandates.

Lesson 1 • Leading Organizational Change for Ethical AI

Applies change management principles to enterprise-wide ethical AI adoption, including resistance management. Prepares senior practitioners to champion ethics initiatives across organizational hierarchies.

Lesson 2 • Ethics by Design in AI Development

Integrates ethical requirements into product roadmaps, design sprints, and engineering workflows from inception. Prevents ethics from being treated as a post-hoc compliance exercise.

Lesson 3 • Measuring and Reporting Ethical AI Performance

Develops key performance indicators, ethics scorecards, and external reporting frameworks for AI governance. Enables organizations to demonstrate accountability to regulators, investors, and the public.