Option 1: Unpacking a key AI concept or concepts

AI is a black box or, at least, that’s how it is first viewed by most people.

It conjures fear in many, simply because its mechanisms are unknown and seem magical (Arthur C Clarke’s Third Law: Any sufficiently advanced technology is indistinguishable from magic).

Most kids in the year 3-6 bracket will have experienced facial (or object) recognition. Using a tool such as Seeing AI will introduce it to them. This tool will speak out a description of a scene or person, will read text, bar codes and other visuals.

How does it do it?

If the demo is interrupted by switching off network access, it will fail. There’s a handy reference to Digital Systems here—clearly the app needs access to another computer somewhere to do the job. If the computer in the phone can’t, is it because it’s not powerful enough? A reasonable, and true, assumption, (for the moment!)

Students can be asked to suggest ways that an object could be recognised: have each student sketch a common object such as a bicycle and ask their partner to see if they can recognise it. A bit like the game Pictionary.

Going to Google’s QuickDraw, will show what over 100,000 people have sketched online when asked to draw a bicycle.

What are the essential elements of bicycleness? Some good discussion should emanate from this and students can try the drawing process themselves online.

A caveat: some school jurisdictions block quickdraw (Grr) so make sure you test and request it be whitelisted first.

So why does the computer need over 100,000 drawings to guess what you are drawing?

There’s a very handy reference to how humans learn here

When parents speak to babies, their language is quite different to normal speech. They repeat words: “Look, there’s a dog. Can you see the dog? It’s a nice dog isn’t it? Do you like that dog?” and so forth. (psychological discussion here)

One of the nice things about AI in the classroom is that it helps kids understand how humans learn!