Discussion
Why I'm not afraid of some future Artificial Intelligence that is super human, malevolent, and doesn't care about humanity, but I am very concerned about the present.
Consider such an AI. We can think of it as being composed of 2 components:
1) tech similar to current AI technology but more advanced (capability part)
2) true, human-like intelligence (life part)
The supposed fear is that the "human-like intelligence" component might not have humanity's best interest at heart.
continued...
The supposed danger, is that the "life part" might be malevolent, self-serving, and an existential threat to human life.
Maybe it's more than a clump of cells on a silicon wafer. Maybe it's a whole human brain in a tank.
Maybe it's not a brain in a tank, but a floating, disembodied human head, Futurama style. Maybe it's a whole "human" in a glycerol bathtub.
The problem is still that this brain might be malevolent, self-serving, and prioritize its own well-being over billions of human beings.
How much of the "life part" is silicon versus organic chemicals, is irrelevant. What the interface from the "life part" to the "capabilities part" looks like, is irrelevant.
The only thing that matters, is that the "life part" sees itself as separate from humanity, is self-serving, malevolent, and controls the "capabilities part." That might lead it to make decisions that put humanity at risk for its own gain. Like destroying the climate, risking nuclear war, or exploiting billions of people.
So my view is that whatever worst case scenario you could imagine under malevolent artificial intelligence, exists today. The "life part" is fashy billionaires. The "capability part" is today's crop of ML systems. For most of humanity, we're already there.
Fashy billionaires already don't make decisions that prioritize humanity. The only thing that would change if malevolent AI developed on silicon, is that ~100 fashy billionaires would suddenly find themselves on our side of the fence. 🤷🏿♂️
Install bonfire.cafe
Get the full app experience
You can imagine a future world (with no medical ethics) in which the "life part" is a human-like coprocessor, which might even be constructed of human brain cells grown on a silicon wafer.😮🤢
That clump of cells interfaces with, interacts with, and controls, the "capability part."
That clump of human brain cells, might not think of itself as part of humanity. It might be completely self-serving, and make decisions that prioritize its own comfort and well-being, over billions of human lives.