That illustration above is what smart people are so terrified of. It is the reason that Americans hate technology and are so willing to rebel against the status quo. The illustration shows an artificial intelligence (AI) mechanism that is self training and ultimately capable of putting you out of a job. Elon Musk and Warren Buffett both have said to be afraid — very afraid that this schematic could destroy the world.
Like we need something else to destroy the world.
I, on the other hand, don’t agree.
The diagram simply shows that (from left to right) data from a source (input sensor) is monitored by a state detector and alerts the machine under a given set of conditions or does something else, like double check the source. Then, if it is important, it does something else. As an example, sensors detect smoke and flickering light and since fire is a threat (or not) and could be very important, the machine calls the fire department.
Now, not all humans can accomplish this and thus this machine save lives. It doesn’t sleep, it doesn’t get drunk and it is always home. Of course its big weakness is that it becomes annoying when the battery gets low and since it can’t change its own battery… well you get the picture.
This mechanism is useful to the degree it performs work. It turns out that it is dependable and is always on the job which is better than many humans. It doesn’t cost much on a daily basis unlike many humans and it can sense many more things than humans can and its always paying attention! Yes, I know what you are thinking. We’re screwed.
But not so fast. There are things we can do that sensors and thus AI won’t be able to do. We can solve problems that have never been solved before. AI needs a history to solve problems. And due to our lazy nature, we can skip thought sequences. If you follow logic which AI does, the machine will end up at a different place than the human who can master the non-sequitur. Traditionally non-sequiturs are a sign of insanity but they are also a characteristic of genius. You will notice this in conversations with the clever in that they will change subjects frequently simply because they have come to the end of a subject in their mind and gone on to the next one while you are still wrestling with the first subject.
In the future sensors will be able to detect human attention and detect the emotional flavorings such as hate and admiration attached to it and that in itself will be very enlightening for humans especially when we discover that these emissions can be weaponized. But, of course, that is another story. This story is still about why AI won’t destroy the world.
The thing that AI cannot do and will not be able to do in the future is empathize. Humans who are more than machines can and can also learn to improve their empathic ability.
Empathy is a number of things depending on how good you are at it but it all boils down to perceiving another space and communicating to that space. It is a gradient form of mental teleportation.
Here are a couple of quick stories as an example. As a teen I never had a headache or any persistent pain. I went to a party and while chatting with a potential girlfriend I developed a headache. After a few minutes of chatting she originated “Pardon me, I have to take an aspirin, I have a terrible headache.” As soon as I realized that the headache I felt was hers, mine went away.
Many years and similar events later I noticed that my eleven year old son was sad. Normally he was a happy carefree soul so this was particularly noticeable. I sat him down and asked him if he was feeling someone who was sad and he brightened up and said yes. One of his friends had a friend who had died from an accidental hanging and that friend was very sad. He originated that he didn’t know the other fellow well. At that point his depression lifted and went off being his happy self.
Empathy is a wonderful thing based upon caring and feeling and can save us from AI destroying the world. Caring and feeling are different than sensing. Sensing is an inflow, caring and feeling are an outflowing. These unique characteristics are expansive and inspirational and are what healthy, growing humans do that world-destroying AI cannot.