AI Systems: Risks, Limitations, and Explainability

Explain how training data, features, and model parameters determine AI behavior and performance.

Identify sources of bias in datasets and explain their propagation through the machine learning lifecycle.

Explain the performance-transparency tradeoff and the technical challenges of explainability in complex models.

Analyze the phenomenon of 'hallucination' in LLMs and its implications for reliability and evidence.

Evaluate AI system vulnerabilities, including 'adversarial attacks' and 'catastrophic forgetting'.

Analyze how choices in data collection, model selection, and deployment create or mitigate legal risks.