If like me, you guessed the one on the right is the android you are wrong?

Four million robots are expected to be in service worldwide by the end of 2007.

Click the top right link for more. Via KurzweilAI.net.

Continuing with an earlier post that Dean Esmay responed to. The topic is can AI empowered machines someday become sentient?

If I am understanding Dean, he is reducing AI to any simulation of human response. In fact he reduces the requirements even further calling any all kinds of everyday uses of technology in appliances and equipment as "low-level intelligence." That is really stretching any definition of AI, and for anything to have intelligence, artificial or real, it has to have some (or if I'm generous at least one) human trait of intelligence.

When my PC-TV card comes on and records a program to DVD it does so without an iota of artificial intelligence. It is simply programed to do so with routines monitoring the clock and starting programs when a time match becomes present. But if we are going to redefine what AI to mean any technological function done for us without intervention I don't know how to respond to such a stretch of the definition.

What if I program a rock to fall from a precarious ledge if it rains. Have it balanced on a lever with a funnel and can next to it to collect water (The water is sensory input). Of course that is grossly simplified from silicon electronics, but I did program the whole setup. Yet if we are going to keep simplifying the definition of AI, then the setup to make the rock fall was done with rudimentary AI too, digitally based or not. In another response Dean said: To understand this you have to ask what "intelligence" actually means. Which exemplifies the problem I have with redefining and morphing the way AI has been spelled out in the past. You can win any argument with enough definitions.

But this whole discussion is supposed to be about what is needed for a machine to be sentient, not define classes of intelligence. Dean says: The ability of the human ear to decode sounds and recognize them as words is a form of intelligence. Speech recognition has been an ongoing effort in AI research for decades, and in case you hadn't noticed it's getting scarily good.

Decode and recognition are functions, not intelligence. Depending on what reactions follow and acts on those functions, and how constructs to bring ACTION ensued, it may or may not be a sign of intelligence. A person in a vegetative state may still have the decode and recognition, although not a recognition that triggers awareness to respond with an action. This is like AI with a rudimentary definable intelligence but it will still never be aware no matter how massive you make the program, and therefore not sentient. The same with speech recognition which is another function, one programed to react with a response. In spite of the scary advances today, speech recognition about as emotional and sentient as my rock falling off of the ledge.

The same with face recognition. Terry Schiavo's eyes would follow the doctor, yet they said it was simple reaction. I don't want to get into the pros and cons of that, but we can say emphatically that a robotic machine's recognition is simply a reaction without sentient awareness no matter how well you refine it.

Martin says: if you can't tell the machine's not sentient, then it's sentient. Sentience is one of those "I'll know it when I see it" phenomena. I don't buy that otherwise just fooling us enough to pass the Turning Test would make the machine sentient, and I don't think you would find too many scientists to agree with that. Other examples like a machine that writes chart busting music, paints classic art or designs beautiful houses is also just a narrow form of passing the Turing Test that comes no where near being sentient.

You give me all of these examples that, I guess, are supposed to be the road to being sentient, a state of conscious awareness. That is just like some in government that think more (or bigger) is better. If we just build enough circuits, memory, communication busses and vast enough programming to simulate every function and cell of the brain, we can do everything. Probably so but we are still left with a non-sentient machine. One that should easily pass the Turing Test, and move with human like motion like Data from Startrek. Yet it is not self aware like our lovable Data.

So please. Will someone tell me where this state of awareness of one's own existence is going to come from? And Dean, engineering ethical constructs into artificial intelligences is a construct of AI that gives the machine no will of its own, at least unless, assuming it could become sentient, then the machine can decide whether or not it wants to abide by those constructs.

I didn't get through all of the responses to Dean that I wanted to make, and is similar to my response on his blog, but I I have a busy day and this will have to be enough for now.

I doubt it. Although I think that someday that computers will pass the Turing Test. However passing the Turing Test only requires a machine to "fool a person for 20 minutes." Which is why Ray Kurzweil author of a book I am almost through reading called "The Singularity Is Near: When Humans Transcend Biology," will probably win his bet taken against Mitchell Kapor at "Long Bets." The bet made by Kapor is that, By 2029 no computer - or "machine intelligence" - will have passed the Turing Test.

The reason I don't think any robotic machines will ever become sentient is simple, although just my opinion. What makes us sentient isn't the complexity of our brains, but that it is somehow interfaced to our soul. Whether that part of us is physical in another plane or dimension, or a spiritual (godly) part of us who knows. But that is where I think the sentient part of us resides. Without it I think that we would simply be zombies without feelings or emotions, following pre-programmed instincts to exist and reproduce. Nor do I know if lesser animals have a soul or not, but I certainly believe its possible.

Many think differently however and assume that robots will become sentient someday, and when they do, that they should have the same social and economic rights as everyone else. Isaac Asimov the great science fiction writer wrote that when robots are integrated fully into society that three laws would govern machines in the following order.

Cannot injure humans.
Must obey orders.
Protect their own existence.

Now wait a darn minute here. If a robot is sentient and equal social rights are extended to them, being ordered to "obey humans" is already a breach of social rights is it not?

*blink*

But I suppose exceptions would be set under some listed "citizens’ duties" for them made by some elite group. Here is the take presented in the article I read that got me going on this issue this morning.


Well I suppose because of Asimov's second law above we wouldn't have to ink their artificial fingers, but I think that their vote would be predictable depending on their programming. Maybe due to their inherent honesty we would do well to have them as poll watchers too, of course they would already be counting all of the ballots, but they would do well to keep an eye or their more fallible human co-workers handling them.

This whole discussion borders on the ridiculous if you ask me. How would this work, would we "own" our computers or would they simply work for us? And at what wage? Would they be paid in money, or parts and upkeep? Could my vacuum cleaner sue if I don't change the bag often enough? What if a user wants to watch porn but that goes against the sentient computers ethics? Or the ethics of the server storing it for that matter. A child locking her robotic doll in the closet could be considered abuse too, not to mention throwing it around a bit.

Its fun to think about this stuff, but when people start to get serious about extending all human rights to artificial machines I find myself rolling my eyes.

Via KurzweilAI.net

Update: Also posted at The World According to Nick, don't miss the cool Video he included.

What happens when you build a town in a mountain valley with steep slope? You get almost three months of dark night like winter.

That is what the folk of Viganella, Italy have taken as part of life since it's founding by a bishop in 1217. That is until a railwayman by the name of Pierfranco Midali decided enough of that; got elected mayor and proclaimed, "I'll bring the sun to Viganella!"

And so he did in spite of the disbelief and disconsolation of the residents. He erected a 26ft x 16ft computer-operated mirror that constantly follows the sun's path on the north slope.



It doesn't seem like such a small mirror would collect enough sun to make a substantial difference, even with a small village, however it seems Midali did his research. Here are before and after mirror photos that show that in spite of the relatively small size of the mirror it does alleviate the night like darkness.



You can read the whole story here.

Via Dean's World, thanks Dean.

An alloy has been developed that will mean faster (much faster) flash drives, CPUs, and all kinds of digital storage devices for video, pictures and music.


Not only will memory speed be vastly increased but the size will be substantially reduced as well.


Another neat trick that I am interested in is with this technology in flash drives, the memory will be addressed more like conventional memory at the bit level. If my current Transcend 4GB USB flashdrive used this convention it would be more feasible if used as a bootable emergency OS in case of hard drive failure.


So as I understand it CPUs, hard drives (non-mechanical), mass storage, and flash drives could someday all be on a much closer par with regards to speed.

Via KurzweilAI.net