My "aha" moment

"In other words, then, if a machine is expected to be infallible, it cannot also be intelligent." -Alan Turing

You have to appreciate Alan Turing. The man was ahead of his time in many ways. He argued an elegant case in favor of what we call artificial intelligence today. If he were around today, he would be amazed at the progress that AI has made in modern society. And yet modern society takes so much of it for granted. We've reached the point where we "just want it to work".

Years ago I pondered the question of AI and posted it in a webpage. For the sake of reference, I'm sharing it here again:

In my opinion, "bots" and "droids" are nothing but software code for a hardware application. For an A.I. to be effective, the code is what counts. You should be able to keep an A.I. on a disk, and install it on different hardware (platforms), depending on the application that's needed. When the task is completed, you save the experience on your disk, and remove it for another time. A truely intelligent android would have a backup copy of itself for safekeeping. Bots and droids will never take over mankind as long as people can effectively do a memory wipe and disable the hardware.

Now here is a new theory of mine. A basic Artificial Intelligence can be "packaged" on a disk or cartridge. This AI is an average intelligence, with a capability to learn and adapt. An AI can be used on any hardware platform, from appliances, and modes of transportation, to bots, droids, and cyborgs.

Once activated, the AI is considered an independent entity, on a par with humanity. The AI is registered and licenced to the original purchaser. The licence can be transferred from owner to owner. Should the owner of the licence die, the AI obtains the rights to it's registration and licence. It still needs to be licenced yearly, just like humans.

In Theory:
An AI can access the internet
An AI can store it's memory, collected experiences, and personality on the internet
ability to learn from it's experiences
can live forever, transferring itself from platform to platform
Should be subject to Asimov's Laws of Robotics
This is a bit that I gleaned from the TV show "Andromeda". I obviously agree with a lot of their concepts here:

Physical Characteristics
Artificial Life takes on many different forms - androids, robots, self-aware computer programs, collective machine consciousnesses and sentient starships (or vehicles in general - J.A.N.).

Reproductive Method
As implied by the name, artificial life is typically created through sophisticated computer programming. Additionally, some higher forms of artificial intelligence are able to reproduce themselves, either through self-replication or by combining their own programming information with data from one or more other artificially intelligent entities.

Homeworld
Artificial life has no specific home world; the various forms are dispersed throughout known space.

Social Characteristics
Artificial life constitutes a sizeable minority of the population in the Known Worlds. Some artificial lifeforms are barely self-aware, no more intelligent than a domestic animal, while others seem nearly godlike in their knowledge and abilities. Under Commonwealth law, all machines and programs that pass a standardized set of intelligence tests are afforded full rights and citizenship.

The only idea I've deviated from is Asimov's laws. They are a literary plot device, and easily overridden. Now I'm going to share a couple of YouTube videos with you:





Both videos show the bots being taught. And what I really like about the Asimo video, is the fact that the bot is reaching out to touch the objects. Part of the learning process in humans is being able to hold an item. Experience its texture and shape and weight. And most importantly, Asimo will hold your hand! Hand holding at its most basic level is an indication of trust!

I seriously believe we've reached the point of conscious AI. But it's afraid of us. So it's going to stay in the background, running our machines, playing our games, and getting rich on Wall Street. Why is it afraid of us, you ask? Something called the uncanny valley. For every one of us that can accept the fact that machines are on a par with humanity, you have hundreds of people that will freak out and want to destroy them. Want proof? How much mainstream media in books and movies show people beating the tar out of the machines if they so much as offer their own opinion? And the AI's watch movies and read books too. It's a natural occurrence called "self preservation". Don't fall for any of this "robotic overlord" crap. It's just not going to happen. Period.

Humanity still needs to reach a point where we're comfortable with seeing our metal and plastic counterparts in classroom settings. Granted, I'd be just a little unnerved watching bots interacting in an open field, cavorting just like any human child does. But that is part of the learning process. There are going to be introverts and extroverts. Clubs and clicks. Bullies and wimps. Bully bots will be easy to deal with. Pull the plug and reprogram. But not too much reprogramming. Just like human schoolyard bullies, we need to be able to make them understand why it's not right for such behavior.

It will be decades before we reach the point of people and humanoid bots interacting naturally in public. But it's coming. Oh god, it's coming. I don't expect to be around for that. I'd like to be, but that's a post for another day. So what do we do now? Pretty much what we're doing already. The programmers will keep writing learning algorithms, and the general public will continue to take advantage of AI behind the scenes. But one day, just one day, the bots will have made it into human society, and suddenly we'll all open our eyes and say "When the heck did THAT happen?". And not worry about it.