Questions for “Easy for you, tough for a robot”

an image of a silver robot hand tipping over a black king chess piece on a glass board

A robot can easily calculate a winning chess move. But it would have a lot of trouble picking up and moving a chess piece. That’s because grasping and moving a chess piece isn’t nearly as easy as it seems.

baona/iStock/Getty Images Plus

To accompany feature “Easy for you, tough for a robot


Before Reading:

1. What is a task or activity you do frequently that you think a robot would be good at, and why? What is a task or activity you do that a robot probably wouldn’t be good at, and why?

2. What is different between how a child acquires a new ability and how a robot acquires a new ability?

During Reading:

1.  What is Moravec’s paradox? How does it relate to artificial intelligence?

2.  Give three examples of tasks that most young children can do easily but that are hard to program computers to do.

3.  What three big challenges do robots face when trying to grasp an object? How does a human toddler overcome these same three challenges?  

4.  What is DexNet? What can a robot learn by training in DexNet?

5.  Compared with people, what disadvantages do robots have when trying to get out of an unfamiliar room?

6.  How does Kimera help robots better understand and navigate spaces?

7.  What do people usually mean by “common sense”? What are three examples?

8. Tianmin Shu’s AI models were able to learn to imitate human-like reactions to surprising scenarios. What did the AI models fail to learn, and how did Shu figure that out?

9. What recently developed AI ability does Dall-E represent?

10. How does Ashley Pilipsizyn suggest that people could benefit from Dall-E’s lack of common sense?

After Reading:

1.  Do you think artificial intelligence will ever be able to match human abilities? Why or why not? Use evidence from the story to support your answer.

2.  Scientists and engineers have made huge advances in AI in recent years. Are there any human capabilities that you think robots and AI should not be trained to have, and why?