Q1. Scenario: You are building a chess-playing AI. Which type of AI (reactive, limited memory, theory of mind, self-aware) does this represent? Justify.
Reactive (no memory of past games) or limited memory (if it learns from past positions). Chess AI like Deep Blue is reactive; modern ones use limited memory (experience replay). It does not have theory of mind or self-awareness.
Q2. Scenario: A customer service chatbot that remembers past conversations to personalize responses. What type of AI is this? How does it differ from a simple FAQ bot?
Limited memory AI – uses past interactions to improve future responses. Simple FAQ bot is reactive (no memory). Limited memory allows context retention, making it more useful for complex support.
Q3. Scenario: Narrow AI (weak AI) vs General AI (strong AI): Give an example of each. Why is General AI not yet achieved?
Narrow: ChatGPT (only text generation). General: a robot that can cook, clean, drive, and have philosophical conversations. General AI requires integrating multiple cognitive abilities and common sense reasoning, which remains unsolved.
Q4. Scenario: An autonomous vehicle must predict human pedestrian behavior. This requires "theory of mind" AI. Explain what that means and why it's challenging.
Theory of mind AI understands that others have beliefs, intents, and knowledge different from itself. It's challenging because modeling human mental states is complex and probabilistic; current AI lacks true belief modeling.
Q5. Scenario: Classify the following AI systems into reactive, limited memory, or theory of mind: (a) spam filter, (b) Tesla Autopilot, (c) a robot that asks "how are you?". Justify.
(a) reactive – uses static rules. (b) limited memory – uses sensor data from last few seconds. (c) limited memory – remembers user preferences but no real understanding of emotions (not true theory of mind).