Thread by @VEverettmfl

Can AI help plan and resource science lessons?
Let's ask it.

It starts with Newton's 1st law. It explains first and then does the experiment. Which is the modern approach. You wouldn't want pupils to think for themselves. It doesn't deal with key terms well at all. Mass/speed/velocity/acceleration. And it completely fails to engage with

It completely fails to engage with the main misconception that if you push things they move and if you stop pushing, they stop.

It has cobbled together things from the internet that all need considerable teacher expertise to make workable for making clear the key concepts.

So all the obvious and easy stuff. But none of the stuff that actually matters to making it successful.

Then the 2nd law. Again, no idea that the fundamental concepts of mass etc are... fundamental. The demonstrations suggested would show that if you push stuff it moves. With maybe some idea that how heavy it is (?) and how hard you hit it makes a difference. Which misses the main

Which misses the main point to address which is the relationship between force and acceleration, NOT force and speed.

Newton's third law. It's more or less given up now with any attempt to get at precise concepts and how to get pupils to engage with them.
It leaves the explanation of the rule to the teacher and suggests some experiments whose purpose isn't clear. Again, it fails to address the

It fails again to address the main misconception which is that the balloon moves because the expelled air is somehow pushing against the stationary air or the walls of the room.

So yes, it can come up with a gloss of the laws and a mishmash of demonstrations.tewchers could use. But a science teacher wouldn't need that.

It has none of the useful stuff that is needed to make it work: engaging with concepts and misconceptions, moulding how pupils think.

But my complaint isn't with this example of what AI can sort of do.
It's the whole concept of:

"Hey! Let's save some time on planning lessons by using a technology that randomly sticks stuff together that it's found."

It's a bizarre dereliction of duty.

Government are permanently on our back to do things better. With NO IDEA of the actual expertise we have. And they think this random mishmash of received ideas compilation technology is an improvement!!!

Going back to the first lesson. The car on the ramp idea is very badly muddled.

It's mixing up so many things:
negative acceleration because of the force of friction from different surfaces - no ramp needed.

Acceleration under gravity - surface friction ignored.

And possibly

And possibly seeking to use the angle of the ramp to balance out friction with the effect of gravity, to measure constant speed in the absence of unbalanced forces.

A great example of AI cobbling together stuff that looks good with no idea of what needs unpacking.

@threadreaderapp can you unroll this thread for me. I seem to have made a mess of it linking from top to bottom. Thanks.