Bring on the (Utilitarian) Robot Apocalypse!

Humanity is flawed in a number of ways. We naturally and inevitably grow weak and infirm. We die long before most of us have gotten all that we want to out of life. We are often cruel, selfish, or careless. We choose short-sighted actions that produce more harm than good. We deplete the planets' resources, and cause a significant amounts of damage to practically every ecosystem on the earth.

We (mostly) all agree that these are problems that are worth solving. And we are making some amount of progress toward addressing some of them. Medical technology has greatly extended our lifetimes, and the promise of genetic engineering suggests that far greater advancements may come. It is not implausible that we could use such technologies to also make ourselves less cruel, selfish or careless, and consequently to help end war, poverty, and needless human suffering. Furthermore, we may have some hope of producing food and energy in cheaper, more environmentally friendly ways.

While it is conceivable that we will address these problems in the next 100 years, it won't be easy. It is difficult to mold one thing into another. We were designed by nature, and nature didn't build us to these problems easily solvable. I suspect that there may be a much simpler way.

Suppose that in the next 50 years, we are able to make and perfect artificial intelligence. There is no clear reason to think that we shouldn't be able to do with chips and software and planning what nature did with neurons and random chance. Suppose further that we are able to create intelligences with whatever sorts of properties we like. We can make them intelligent or stupid, emotional or stoic, we can give them a sense of humor. Etc.  If we can make succeed in making malleable artificial intelligences, then we can also make artificial lives. Such creatures would get the same sort of value out of their lives that we do out of ours.

It strikes me as extremely plausible that we should be able to more easily perfect AIs than perfect human beings. We can design AIs from scratch, and give them whatever features we want. They won't have to worry about war. There is no reason to think that they would have to worry about aging either. Since they would not have the same natural history as we do, they would also be immune to disease. They could, plausibly, survive outside of the earth, and could plausibly be sustained much more easily in an environmentally friendly way.

In order to perfect human beings, we have to tamper with an existing thing. It might be possible to get rid of human flaws, but probably at a greater effort and cost.

So if we ever get to the point where we can develop such AIs, why not just put the human race to rest and allow these AIs to take over? We took over from homo heidelbergensis, and they gave birth to us. Why not do the same thing with artificial intelligences. What do we have of value that we can't program into an AI?