Artificial Intelligence Lesson 4 of 4
AI Safety, Bias, Hallucination Prevention & Ethics
18:00 AI for Everyday Problem Solving
Learn defensive prompting strategies to guard against hallucinations, prompt injections, and data privacy leaks.
Lesson notes
Deploying AI responsibly requires understanding where models fail. Hallucinations occur when a model encounters high uncertainty but is prompted without grounding citations.
What you'll learn
- Implement Retrieval-Augmented Generation (RAG) grounding
- Prevent prompt injection attacks in user-facing inputs
- Understand regulatory and data privacy considerations
1. Grounding Prompts with Strict Boundaries
Always instruct: 'Answer the question using ONLY the provided reference documents below. If the answer cannot be found in the text, state clearly: I do not have enough information to answer.'