What is the logical progression of human relationships with intelligent machines?
The short answer:
Humans react to machine intelligence with the same behaviors they do to other forms of intelligence.
When we don’t understand the other form of intelligence, such as animals, we exploit them, often to the point of their extinction and our detriment?
When the other intelligence is human we behave instinctively. We socialize, compete, and exploit even exterminate to satisfy instinctive drives that have evolved to benefit our genes and our species.
There are many examples of this behavior. We respond to mammalian pets with human responses. We love them, protect them and provide for them as we would children. This is because the closer we relate to mammals the more we see them as intelligent. But do we treat pet fish in the same manner? No. We capture them and keep them as curiosities. We neither recognize their intelligence or respond the same way we do to mammals.
How do we respond to machines? Historically and still currently we treat machines as inanimate objects. They are more fish than mammals to us. We need our tractor on the farm. We maintain it and exploit it for our benefit. But we don’t consider it intelligent. Sooner or later it will be thrown on the scarp pile.
But this relationship is changing. A tractor (for the moment) has no brain, no mind. It doesn’t speak to us. It doesn’t, smile, bat its eyes or wrap its arms around us. But it soon will. Science is striving to create machines that look and behave, (mimic), human behavior. Let’s ask the question: Why?
We have come a long way with inanimate machines. Why strive to duplicate humans in machine form? At this time, I have, too many answers to this question. For now let’s just say (we pursue it to prove we can do it). I will give you one reason. We do it because our enemies will do it. And in the mean time we will use it to better exploit our fellow humans.
But is this really dangerous?
We have reached a point in the development of AI where we don’t understand exactly what’s going on inside the machine to achieve its output or action.
AI is a non deterministic process. This means that for the same inputs it can exhibit different behaviors. This means it is unpredictable and possibly dangerous. Imagine if a nuclear reaction were unpredictable. Start the reaction and it might provide clean abundant entry. Or it might start a chain reaction and destroy the Earth. Would we take the risk? The answer is yes?
There is a famous story of Robert Oppenheimer asking Author Compton if he thought that the first nuclear bomb might start a chain reaction that would turn the Earth into an expanding fireball. They reasoned, (not without great debate) that the enemy, Nazi Germany, would most likely develop and use the weapon. They pressed ahead and eventually exploded the first atomic bomb.
There are countless example of this same reasoning being applied to new potentially dangerous research. We will continue to develop AI with the risk of disaster because it is in our nature to explore and to exploit and because our enemies will do so.
We never put genies back into the bottle.