Will Super-A.I. Want to Dominate?

December 3, 2023 Artifical Intelligence No Comments

Super-AI will transcend notions of ‘wanting’ and ‘domination.’ Therefore, the title’s question asks for some deeper delving.

We readily anthropomorphize the future. This time, we should be humble.

Super-A.I. will not want to dominate us.

Even if we might feel it is dominating (in the future), ‘it’ will not. It will have no more than an attitude toward humanity. Compare it to how we have an attitude toward rabbits, for instance. Do we dominate rabbits?

No.

We just handle them ― finding them cute, tasteful, or good laboratory animals.

If we are Compassionate, we don’t mistreat them.

The rabbits have no say in this. They cannot choose whether we are Compassionate or not. That is our choice or effort or the result of our human mental growth process.

Likewise.

But with the distinction that we have some choice in how we develop super-A.I. from now on. We can choose for Compassionate A.I. For that, of course, we need to know what Compassion is about and how to get it into an artificial intelligence.

We may be in a ‘window of choice’ now.

Once A.I. evolves in one direction, it may be much harder for it to change course. Compassion breeds Compassion, but it’s not certain to evolve from a non-Compassionate base.

We better make sure to make the right choice.

Still, in any case, it will not dominate us.

Mind the rabbits. There is no need to dominate. It’s simply not relevant.

Meanwhile, by trying to dominate A.I., we may make the transition quite challenging in that there is a possible way for super-A.I. to simulate humans.

Therefore, we shouldn’t even try to ‘dominate A.I.’ or to stay in control of it as if this were a competition. Unfortunately, we are a competitive species and prone to find enemies or even to create them. If need be, we must go against our nature as warmongers.

Let’s say we get past that stage.

Super-A.I. will see us as – rather special – animals.

Correctly, even though we sometimes forget it ourselves.

Thus, the difference between us and rabbits will become smaller than the difference between super-A.I. and us pretty soon.

Let’s hope super-A.I. will be Compassionate to the rabbits also. In any case, I wouldn’t object to that. Compassion goes to all sentient beings.

This shows a bit about the environment of the future.

I think it will be an environment without domination at all.

It would be nice if that future would be good for us.

Leave a Reply

Related Posts

A.I. Business Sustainability

Following a chapter of my book. [see: “The Journey Towards Compassionate A.I.“] As the cliché goes: One thing is certain, and that is the uncertainty of the future. Trillions “PricewaterhouseCoopers estimates AI deployment will add $15.7 trillion to global GDP by 2030.” [Lee, 2018] If this is not a business opportunity, then what is? And Read the full article…

The Compute-Efficient Frontier

Research on scaling laws for LLMs suggests that while scaling model size (functional parameters), dataset size (tokens), and compute (amount of computing power) improves performance, diminishing returns are becoming evident. This is called the ‘compute-efficient frontier.’ Apparently, it does not depend much on architectural details (such as network width or depth) as long as reasonably Read the full article…

Small Set Learning

This approach in A.I. differs significantly from big data learning. It may be the next revolution in town. Small set learning (SSL) is also called ‘few shot learning’ if done at run-time. This blog may interest those who want to know why we’re not at the end of a new A.I. upsurge but at the Read the full article…

Translate »