Michele’s Freeminding

I’ve always been interested in AI. And I’ve always wondered wether it is possible or not to create the perfect AI.
The kind of AI that can understand, thus emulate human emotions.
And I’ve realized it’s impossible.

It’s impossible because to recreate emotions you should be able to understand emotions and systemize them, which is IMHO impossible.
Why? Because it would mean understanding the essence of the human being.

What’s so special about us as human beings is the uniqueness that is inborn in each one of us. This is why it’s impossible — or at least, unlikely to happen — to elaborate equations that will allow to build a perfect model.

We can aim at creating likely human emotions emulators, but in the end we will never reach perfection. We’ll probably get very close, but it will always miss that something that would make it perfect.
And that something is explaining the irrational part of each and every person.

And not only it’s, to me, impossible to modelize something irrational as it would be rationalizing the irrational, but also it’s impossible to make it wide enough to include every human being ever existed/existing/that will exist — remember everyone is unique.
We could build a statistical model to portray reality in a decent way, but still it wouldn’t be enough.

Probably one day we’ll be able to elaborate a human-enough AI, but it’ll always remain a lame attempt at emulating reality.

This post was written 2 years, 7 months ago on December 29th, 2005 late evening.