12 · Final quiz · ElementaryMBA

1. When you add AI to a robot, which part of the sense → think → act loop mainly changes?

The sense step — it needs new sensorsThe think step — it can use a model that learned from examples instead of only typed-in rulesThe act step — it needs new motorsNothing changes

2. What is the difference between programmed rules and learned behavior?

Rules are written by a person in advance; learned behavior finds the pattern from examplesThey are exactly the same thingLearned behavior is always more predictable than rulesRules can only run on Mars rovers

3. A robot uses 'computer vision' to do what?

Charge its battery from sunlightMake sense of images and video — like recognizing a stop sign — by matching learned patternsStore its programTalk to other robots over radio

4. Why might a voice-controlled robot understand the person who set it up but struggle with a friend?

Robots can only ever hear one personThe model learned from examples that may have missed other accents or noisy roomsMicrophones stop working for new peopleIt is choosing to ignore the friend on purpose

5. Why does a Mars rover need to drive itself, when a sidewalk delivery robot could be watched by a human?

Mars rovers are cheaperMars is so far away that a person can't steer it in real time, so it must make its own routine decisionsDelivery robots have no sensorsRovers don't actually move

6. In a tool like micro:bit CreateAI, how do you teach the robot a new skill?

Write hundreds of if/else rulesCollect labeled example movements, then train, test, and improve a model on that dataBuy a more expensive boardWait for the robot to learn on its own with no examples

7. What is 'sensor fusion,' and why is it useful?

Melting two sensors togetherBlending several sensors so the AI makes one better decision, covering each sensor's blind spotsTurning a sensor into an actuatorUsing only one sensor to save power

8. For the highest-stakes action a robot can take, what do careful engineers often do?

Leave it entirely to the learned AI's guessUse a plain, readable hard rule (and/or keep a human in the loop), because a learned guess can be confidently wrongRemove all the sensorsLet the robot decide with no testing

9. Which is a fair concern about automation and jobs?

Robots never change any jobs at allAutomation can remove some jobs while creating new ones, so we should help people whose jobs change and keep new opportunities open to everyoneOnly robots can have jobs nowAutomation always helps everyone equally with no effort

10. When your trained model keeps confusing two movements, the engineer's fix is to…

Give up — the model can't improveIterate: add better, more varied training examples and test againWrite a thousand if/else rules insteadBuy a faster computer and change nothing else