The station telescope clicks all night. By morning, mission control has a stack of four hundred snapshot cards.
"Every card needs a label," Raven reads from the message. "Star, planet, comet or satellite streak."
Rocket fans the cards out like a giant hand of a card game. "We could sort these for a week and still not finish."
Nova drifts over the pile, her blue lights pulsing slowly. "Mission control wants a sorter," she says. "They are calling it Sky Sorter."
Raven frowns. "Would Sky Sorter count as AI, or just a fancy checklist?"
"That," Nova says, "is exactly the right question to start with."
Nova is the crew's AI companion drone. She is a story character, so she can do things no real machine can do.
Real AI is different. It is a tool that people build. It does not think or feel like a person, and it can be wrong.
This course is unplugged. You will use paper, pencils, index cards and tables, never a real AI tool.
Scientists at a United States standards agency wrote a careful definition of an AI system. Here it is in plain words.
An AI system is an engineered or machine-based system. For a set of objectives, it produces outputs such as predictions, recommendations or decisions.
AI systems also work with different levels of autonomy. Autonomy means how much the system acts on its own.
| Part of the definition | What it means | Sky Sorter example |
|---|---|---|
| engineered or machine-based | People design and build it | Mission control designs the sorter |
| a set of objectives | The goals it is built for | Label every snapshot card correctly |
| outputs | What it gives back | This card is probably a comet |
| levels of autonomy | How much it acts on its own | Does a person approve each label? |
| Statement | True or false? |
|---|---|
| An AI system is built by people for a set of objectives. | ? |
| Every AI system acts with exactly the same amount of autonomy. | ? |
| Nova is a story character, not a real product. | ? |
| Real AI thinks and feels the way a person does. | ? |
Sky Sorter is our running project for twelve weeks. It is a paper sorter, built from tables and rules.
It sorts telescope snapshot cards into four labels: star, planet, comet and satellite streak.
On the crew's cards, a star is a tiny sharp point and a planet is a small round disk.
A comet card shows a fuzzy glow with a tail. A satellite streak card shows a thin straight line.
You will design its features, train it, test it, find a bias in it and finally review it.
Strong start, explorer. Tomorrow you will sort outputs into predictions, recommendations and decisions.