AI and the Golden Rule

Better be nice to AI so that when it takes over the world it's nice to you back.

I've heard people say variations of this dystopian (utopian?) hype for a long time. 

But when you break the concept down, it's actually something that's already pulsing through our online world. 

Generative LLMs learn from us. They scrape everything on the internet and use it to predict answers to our queries. 

If it finds nice stuff on the internet - boom - output of nice stuff.
If it finds mean stuff on the internet - boom - output of mean stuff.
(Of course it's finding a mixed bag online, and I'm also not accounting for the guardrails that are implemented to keep responses "nice"). 

But the point is, it's learning from us. 

So when people say "be nice to it so it's nice back"...it's less of a dire dystopian warning and more of a golden rule do-unto-others type statement. 

And this is a benevolent double bind, because if we populate the online world with interactions that lift one another up then two things happen:
1. AI starts learning more of the nice stuff from us and using it in outputs
2. We rewire our cognitive and habitual landscapes to practice patience and tolerance toward each other
(both of these outcomes are positive upward spirals)

So maybe the moral of the story is to be nice to AI because your interactions with it are shaping you and shaping all of us.