Q1. Scenario: You are designing a chatbot for a customer support system. How would you define Artificial Intelligence (AI) in this context, and what capabilities would the chatbot need to demonstrate to be considered "intelligent"?
AI is the simulation of human intelligence in machines that can perceive, reason, learn, and act. For a chatbot, it needs natural language understanding, context retention, ability to answer queries, learn from interactions, and handle exceptions, mimicking human conversation.
Q2. Scenario: A self-driving car must decide between hitting a pedestrian or swerving into a wall. How does AI enable such decision-making, and what type of intelligence does this represent?
AI enables decision-making through sensors, perception, and planning algorithms. This involves reactive and deliberative intelligence: sensing environment, predicting outcomes, and choosing actions based on predefined ethics or learned policies (reinforcement learning).
Q3. Scenario: You are building a system to recommend products based on past purchases. Is this AI? What is the difference between a rule-based system and a learning-based AI in this case?
Rule-based system uses fixed if-then rules (e.g., if bought milk, suggest cereal). Learning-based AI uses data to discover patterns (collaborative filtering). Both are AI, but the learning-based adapts and improves over time, handling complex relationships.
Q4. Scenario: A medical diagnosis system uses AI to analyze X-ray images. What are the key components needed? Relate them to the definition of AI (acting humanly, thinking humanly, acting rationally, thinking rationally).
Components: perception (image processing), reasoning (pattern recognition), learning (training on labeled data), and action (output diagnosis). This aligns with "acting rationally" – choosing the best diagnosis to maximize correct treatment, not necessarily mimicking human thinking.
Q5. Scenario: A voice assistant like Alexa or Siri uses AI. Give two examples of where it uses "narrow AI" and where it might need "general AI" in the future.
Narrow AI: speech recognition and weather query answering. General AI: understanding context across domains (e.g., planning a trip based on calendar, email, and preferences without explicit commands). Current AI is narrow.