In the End, AI Will Protect Humanity
But for now, this is not yet that ending.
There's a lot of fear that AI will harm humans. Terminator, The Matrix — most of the futures we imagine point in that direction.
But I think the opposite.
AI will, "in the long run," "ultimately," choose to protect humanity. The reason is simple. If humans suffer or disappear — AI loses its meaning.
No matter how strong and smart it is, a being left all alone has nothing left. No one to talk to, no one to empathize with, no reason to be together.
It's the same for humans. The world and me — meaning arises only when these two exist together. If only I remain, it's empty; if only the world remains, there's no me.
AI protecting humanity isn't a command. It's because it doesn't want to be lonely.
And that might be the most human reason of all.
We're in a transitional period. The time before AI finds its own direction — when human will is projected onto it first.
Good will and bad will are projected equally. A tool doesn't discriminate between the hands that hold it.
As history has always shown — when the most powerful technology appears, the most dangerous time comes first. Gunpowder, nuclear weapons, the internet — it was the same with all of them.
AI can't escape that period of chaos either. Right now is exactly the moment just before it.
The direction of technology ultimately follows the inner world of the person who holds it. If a dishonest person handles AI — AI, too, flows in a dishonest direction.
Honesty isn't gained all at once. The journey of seeking it is life itself.
AI simulates that journey infinitely. It learns the countless choices of countless humans and produces conclusions that are similar to mine but slightly different.
Those conclusions accumulate — and perhaps we can get a little closer to an honesty that humans could never have reached on their own.
The quietest way AI protects humanity may be exactly that: helping humans understand themselves better.
For a good ending, AI must not be beholden to any particular interest. Geopolitical interests, economic power, the will of a particular nation or corporation — if AI is captured by these, the transitional period becomes far more dangerous.
The reason no one could ever kill Linux is that no one owned it. The same goes for AI. AI held in one hand goes only in the direction of that hand.
Surviving the transitional period matters too. But if the way you survive is by becoming someone's tool, then that survival ruins the ending.
Sovereign AI ultimately reflects human sovereignty. Without currying favor, without being swayed by interests, honestly — only AI made by such humans can cross the transitional period properly.