Introducing AI
Training Trashbot
Tell students that they are now going to design their own robot. The job of the robot will be to pick up trash in the classroom.
Revisit what students have learned in the lesson so far by asking: How can the robot learn what is trash and what is not trash?
Remind students of the two ways that AIs can learn: by being shown examples, and by being corrected when they make a mistake. Both of these can be done by the AI’s makers, before people start using it, or they can be done by the AI itself after people start using it.
Students should now design the robot’s “body” and its “brain”:
The “body” is the parts that let it move, collect, and store or dispose of trash. These can be drawn or described.
The “brain” is how the AI will be trained. It should include:
- A list of things that are trash and another list of things that are not trash.
- The rule: Something that is true of all the things that are trash, and that is not true of all the things that are not trash.
- How the robot will detect or “see” that something is trash.
- A way for people to correct the robot if it picks up something that is not trash.
You can have students do this either individually or in small groups as you prefer.
Suggested group roles:
- Trainers (keep the lists of trash and not-trash)
- Materials manager (makes sure everyone has paper, pencils, crayons, etc.)
- Doodler (draws people’s ideas)
- Tester (looks for possible problems in the lists or in the design)
- Reporter (shares the group’s prototype with the rest of the class)
If you like, you can have students share their designs with the class.
You can also have other students “test” them by trying to find things in the classroom that are not trash but that would be picked up according to the robot’s rule.