We are going to create AI that matches general human learning abilities soon. This means we will have machines that can learn just like a child does, and can learn to create and solve problems on their own, and learn any subject and skill, a human can learn.
We will then create AIs that have larger brains that learn with higher resolution, and more memory. So think of a machine like Google or Wikipedia, but which doesn't just index and store billions of pages on the internet, but has read and UNDERSTANDS the data in all those pages. So you can ask it a question just as you would ask an equation of a human expert in a given field. But the AI experts will know more and understand their subjects better than any human ever could. These machines will be the tools we use to learn about anything we need to learn about.
In addition, the general AI machines can be configured with any combination of sensors and arms so it can be optimized to do whatever job it is built to do far better than any human could. Self-driving cars will have “eyes” looking in all directions, not just forward like humans. It will look backward and to the sides as well as forward at all times. But unlike our current self-driving cars, our future AIs will have a deep common sense understanding of what it is seeing that our current self-driving cars do not. A kitchen cook-bot might have 10 arms and many eyes so it can be doing the work of 5 humans at once.
We will fill our lives with these smart machines just as we have filled our lives with engines, lights, phones, electric motors, and computers. All the machines will be smart in different ways and they will be built to satisfy our needs.
The smart machines will have the mental ability of humans, but not the motivation to be selfish and greedy with only their own needs always front and center like humans. They will be built to have the needs of the humans they work for as their top priority. So they won’t be this threat that the AIs will enslave us. We could build selfish AIs that would be a risk to our lives, but for the same reason we all don’t have nuclear weapons for home defense, we won’t live in a world of dangerous AIs either. We will live in a world of super-smart machines that take care of all our needs.
How smart an AI can be will be limited only by the amount of computing power we can justify spending to create it. We will build a few very big super-AIs but the world will be mostly filled with small not so smart machines helping us.
We will use the smart machines to explore the universe for us and report back to us what they found out there. We will use smart machines to do all the science and engineering for us. They will figure out the mysteries of the universe for us and then train us on anything we want to learn (to the limit a human can learn it).
Today, we are used to living in a world where we are the smartest machines on the planet. In the near future that will no longer be true, but we will still be the reason the machines are here. They will work for us, not take over. But this won’t be hard for us to deal with any more than the fact we don’t mind that a calculator can do math 1000x times better than we can. I don’t try to calculate logarithms in my head when I can just pick up a calculator to do it. I don’t have to be an expert on much of anything when google will always lead me to the facts about anything I want to know, and in the future, we won’t just have a google leading us to pages written by humans, will have access to endless AI experts on anything we want to know more about. Medical advice, shopping advice, current social trends advice, scientific experts, history and economic experts AIs. We will all have AI experts at our fingertips to tell us anything we want to know about anything that is known in the world.