TALAMANA · THE AI LITERACY MAP FOR ARCHITECTURE AND DESIGN · Foundations · AGE 11—15 · FACTUAL · DURABLE
Everything it knows is data
To the machine a picture, a word and a song are all numbers that people collected.
The idea
A machine does not see a picture the way you do. To the machine, the picture is numbers. So is a word. So is a song. Everything a machine "knows" is data that people collected and gave it. Its world is that data, plus whatever you give it in the moment. Nothing else exists for it.
Why it matters
If the data is the machine's world, then whoever chose the data chose the world.
See it in the studio
A machine that has only seen data about city houses knows nothing about your village house. It is not being unfair. Your house is simply not in its world.
Watch for this
Saying "the machine knows everything." It knows its data, and its data is never everything.
Try it
Describe your school to a friend using only ten photographs. What did the ten photos leave out? Every set of data has a gap like that.
Prove it
Explain why a machine can only answer from what people gave it, and give one example of something it would miss.
How it works
Collected examples used for teaching machines are called training data. Later ideas on this map — where data comes from, who is in it, who is missing from it — all grow from this one. Later still, you will learn that a machine can also be handed data while it works: a photo you upload, a document it looks up. That is a different channel from the training data, and a card higher up this map keeps the two apart.
What this idea builds on
- A starting idea.
What this idea opens up
Sources
- mt_kH1DzOPsXG (Marble Skill Taxonomy, CC BY-SA 4.0)
- TEC.U
- Elements of AI
Open this idea on the map · The complete map · Logika · RBDS AI Lab, India · revised every edition.