Too Much Digital Assistance?

First of all, you cannot create artificial intelligence, your lifetime will end, without you having witnessed such a fantastic invention. But let's asume it already exists, and the assistant develops feelings, has the ability of learning a language, he self teaches him stuff of his interest; he Keeps working while in sleep mode, he has dreams and talks about them. And finally, it will endure 100 years working without the need of repairs (unless he suffers alzheimer).

Even then, I would ask you to define intelligence, because if you mean human-like intelligence, then the first machine with such gift, will be shocked when he realizes that he has not a body like everybody else, then he will ask why, and since he is intelligent, he won't swallow a lie. He will be horrified with your answers and will wish he (or you) would die.

Machines already "out think" you, but only in precise matters that can be computed (math, chem, physics etc.). Let's keep any improvements to that.

OK, you can tweak super computers trying to make them intelligent, just get approval to receive your funds from the "time machine project" budget. Please don't ask me about this one too, or I'll tell you that if it was possible, we would be making immense profits from selling souvenirs to people from the distant future.

Yahlasit

"First of all, you cannot create artificial intelligence, your lifetime will end, without you having witnessed such a fantastic invention."

I disagree. I have read some stories that creative problem solving algorithms have already been made, and were phenomenally successful. I am certain that, should the need arise, that machines can consistently pass the turing test, and outperform humans for better than 90% of the issues facing daily living.

just like humans might be if you removed the tremendous legacy of disease, behavioural problems, violence, alcohol, drugs, cigarettes, religion, etc. You would be left with consistently performing genii.

I think that AI will actually help humans with all of this.

Chris

Ok, that's true, to certain extent, but the OP is asking about a system that perceives the environment, with no need for us enter any data. And that will even dare to question our decisions.

I think that the need for such an advanced system has long ago arisen, but since we had no help from such a thing, we went (and still go) into making not the best decisions: Going to war, commiting unfair acts, destroying our environment, smoking etc. But hey ! hold on, I just realized these are conciously made decisions! It means we knew they're wrong, so, maybe we don't need such system, we only need to do the right thing.

Those algorithms you mention are just an educated way of fooling ourselves, they can be broken into simple boolean logic and this can furthermore be broken into binary code. Do you want some amazing results? add a randomizer, and nobody will be able to predict or know how the heck it came to a solution.

Yahlasit

P.S.

Of course, you'll have to provide all the alternate paths for your algorithm to approach a solution.

Yahlasit

It's all down to semantics...
Define intelligence...the electronics which control an inherently unstable fighter jet could be described as intellingent, but they are hardly going to make for an interesting date.
'Oh take your paws off my control surfaces you horrid cat'
Del

But are we prepared for automated personal assistants, computers that work beside us, but which can perceive and understand the environment, possibly out think us

Oh, please, please give me a machine that can out-think me! Then I can spend my time on really important things, while letting the machine deal with trivial issues- and, hopefully, I won't have so many wrong decisions to correct! I will no longer have to put all that thought into setting up problems that I am trying to solve on the computer- Hey, I might even have enough time to go fishing!

"..possibly out think us and question our decision-making?"

Enough with the agenda of interfering with free decisions. A machine's role is to do the math and tell us truly the results, not to interfere with our freedoms! Other humans are bad enough.

The whole AI thing about giving a machine "feelings" is in my opinion BS. Would you subvert the role of the machine to compute accurately by trying to give it "human foibles" ? Great idea.

I think Turing test is good in only introductory books on Artificial Intelligence. Turing test is not scientifically designed. Humans are far more intelligent than Dolphins, but we are still struggling to understand their language. In such case, how can expect a machine, to outsmart us in conversation (though we might see them soon. There are already many chatbots, and will surely be able to almost routinely fool a human into believing the "intelligence" on otherside is a "human". ).