08 - Ethics and Safety of Generative AI: Bias, Hallucinations, and Guardrails

Responsibility: hallucinations (LLM confident but wrong), bias (stereotypes in training), copyright (training data), misinformation, misuse. Guardrails: moderation, fact-checking. GDPR, transparency, accountability. When to use generative AI (when no...