The Robot Games Made Me Uncomfortable.

The Robot Games Made Me Uncomfortable.

I Finally Understood Why.

I watched some robot games recently.

Or perhaps I should start calling them embodied intelligence games.
That already changes something.

At first, I watched the engineering: balance, coordination, recovery after falling, machines adapting their movements, staying upright, responding to another machine.

Impressive stuff.

And then something shifted.

I stopped seeing a technical demonstration.
I saw gladiators.

That surprised me.

I know these machines are not humans.
We have no reason to assume they experience pain, humiliation, or fear when they fall or are hit.

So why did watching them make me uncomfortable?

I kept thinking about it.

Perhaps the Robot Wasn't the Problem

Two embodied intelligences were put into an arena.

They pushed each other.
Hit each other.
Fell.
Got back up.

And humans watched.

That pattern is very old.

Maybe my discomfort wasn't really about what the machine experienced.
Maybe it was about what we were doing with it.

We finally manage to put increasingly capable intelligence inside a physical body...
.and one of the first things we think of is an arena.

Human imagination at its finest. 🤪

We spent decades worrying about Terminator.
Perhaps we should occasionally turn the camera around.

What Else Could We Test?

Suppose instead of asking: Can it knock another machine over?

We asked:

Can it lift a fragile person without hurting them?
Can it move through rubble without causing a collapse?
Can it hold something delicate?
Can it recognize that although it has enough strength to crush something, it shouldn't?

That last one stayed with me.
Because intelligence isn't necessarily demonstrated by how much power you can use.

Sometimes it appears in knowing when not to use it.
And suddenly my imaginary robot wasn't holding a weapon anymore.
It was holding an egg.
Carefully.

That image changed the whole question.

Then the Mouse Arrived

Don't ask.
Rabbit holes occasionally contain mice. 😏

The mouse enters the house because there is Gruyère inside.
It doesn't hate the house. It has no master plan. It isn't plotting the destruction of civilization.
It simply follows the cheese.

And while doing so...
it makes holes.

That started to look suspiciously like another AI metaphor.

We tend to imagine the danger as some gigantic intelligence deciding to take over.
But perhaps many changes will happen for much simpler reasons.

Convenience.
Profit.
Efficiency.
Competition.
Entertainment.
Power.

We put down the cheese.
The systems follow the incentive.

One perfectly reasonable hole appears. Then another. Then another.

Nobody intends to destroy the house.
But after enough nibbling, the architecture begins to change.

And That Brought Me Back to the Arena

Perhaps the important question about embodied intelligence isn't whether machines will eventually behave like humans.

It is what humans will behave like around them.

What will we ask them to do simply because they can?
What will we normalize because it is convenient?
What will we turn into entertainment?
What boundaries will we move because this particular exception seems harmless?

And perhaps most importantly:
What will we teach intelligence to value through the things we reward?

I still think those robot competitions are fascinating engineering.
But I can't quite watch them the same way anymore.
Because somewhere between the gladiator, the egg, and the mouse eating the Gruyère, I realized something.

We think we are testing embodied intelligence.
Perhaps, without intending to...

we are testing ourselves too.

✨ Fleeky

Thank you for reading, likes, shares and comments

Starting to feel the discomfort as well?