AI on its way to "think for itself"?

Found this interesting, what next ? It gonna start having emotions ? :grin:

I don't hope we get the same fate as Terminator or Mass Effect :sweat_smile:

People keep saying AI will take peoples jobs, but perhaps there is a new career path as therapists for troubled AI.

Anyone remember Marvin the Paranoid Android, in "The Hitchhiker's Guide to the Galaxy" by Douglas Adams.

2 Likes

I guess a Therapist for it comes too late, hahaha! Maybe a fresh Towel would help ...

"... And then of course I've got this terrible pain in all the diodes down my left side. ... .... ....". Loved everything he wrote :heart:

I'm tired of linking to the YouTube video about A.I. will kill us all. As the speaker explained years ago, A.I. is machine code that writes itself with no intervention or participation by humans, other than base concepts. Once it is fully functional it makes decisions on logic, not on emotion/conscience which is the fundamental difference between humans and machines.

https://ugetube.com/watch/artificial-intelligence-it-will-kill-us-jay-tuck_cHiBi41rfBEpTqO.html

There are a lot of important things to consider on this and before this thread gets too carried away....

It is important to step back away from Fiction and look at the actual science involved here.

To say that "A.I. is machine code that writes itself with no intervention or participation by humans" is a very misleading statement.
Yes, some GANs and autoML (Machine learning) can write code to expand their learning, but this is performed only within the framework - all of which is under oversight from humans, developed as algorithms entirely by humans and human participation includes vetting and analyzing the results.

It is true that A.I. in its current state lacks emotion and conscience. To add to this, it also lacks intelligence.
The best way to describe our current A.I. is that it mimics the appearance of human behaviors (Settling firmly in the uncanny valley) by repeating recognized patterns and then amalgamating the data to produce a logical output.

While this is sophisticated and advanced when compared to the NES game "Super Mario Bros." The principle of output defined by input is the same.
With SMB on NES, the input is the user on the end of a wired controller. With A.I. it is many users having provided data across many mediums over many years.
But the Principle
remains
the same.

There are many different models of A.I.
But one thing they share in common is that unlike Terminators "Skynet", they are not structured in a way that they can "Wake up one day" and be sentient.

That is not how these programs work at all.

Terminator is fiction. It is a movie.
We all know that Movies focus on sensationalism and art. Not on getting the science right.
Explosions making noises in space.
The doofus hero always getting the girl at the end.

Fiction.
Movies are made to please the audience, not to reflect reality.

I often caution readers about A.I.
But it always has nothing to do with it turning into monster killer robots.

It has to do with taking its answers for granted, since it is incapable of independent conscious thought.
It has to do with its lack of intelligence allowing it to be easily exploited.
Humans, the creatures that can experience emotions including greed and unethical desires, can use it as a tool to harm others. And it can offer no resistance because it cannot think about what it is doing.

I read the article link in the O.P. and while interesting, no part of the content actually says that any of the content shows any step of A.I. being able to think for itself.
That is not what the article says, in spite of it beating a drum. Even the click bait title does not say that.

This thread title is what makes that suggestion.

See the human sensationalism telephone game at work.

1 Like

I didn't suggest it thinks for itself , I said "on its way?"

And I wouldn't be surprised if it does actually starts doing that in a couple of years. But that's just me lol.

You quoted me. I did not say you said or suggested it thinks for itself.
I said:

That said, the article does not suggest it is showing any step toward nor on its way or anything similarly worded.
Instead, it focuses on what it can currently do in its current parameters and that one researcher was surprised how this is similar to a function that many mammals can do.

And this is actually unsurprising.
A.I. models create output based on the amalgamated data of what human responses would be.

My largest concern with AI, is the abuse of it, by humans. Humans have an uncanny ability to do whatever is necessary, if the end result, benefits them. It doesn't matter what is moral or ethical, if a human can use AI to make money, they will.

The abuse of AI is already being used, to either sniff on people's information, or to steal other people's photo's, and then push something out, and they call it AI art. We already have laws in this country called copyright infringement and plagiarism.

But somehow, AI does it, and gets away with it. The issue is AI, got spearheaded so fast, there were no laws put in place for the use of it. But now our government has been made fully aware of the dangers now, and are in a slow process to write new laws, to prevent the missuse of AI.


Maybe because it's too OTT... just because some people believe "AI will kill us all" doesn't mean it will, or that everyone will be persuaded by watching the video.

AI is dangerous but it is also very useful. At the moment it's still way to unreliable and fallible... it is, however, a way of harvesting data and making money. The humans feeding off AI's profits have zero conscience either. It's always been about money and power.

Humankind is more likely to be taken out by the Yellowstone caldera any time soon than by AI running amok, and there doesn't seem to be a whole lot of logic, more hallucinations and trying to come up with any kind of accepted answer.

I actually agree with the statement, despite it's lengthiness. AI does lack intelligence, even though AI LITERALLY stands for artificial intelligence. Cause tbh, AI may be smart, but it's not going to just become sentient like PAL from The Mitchells Vs The Machines (which is the last movie I watched before the new year).

In all seriousness, people will think AI will take over the world, but really, AI is just taking itself. Badly.

I see about that. Even schools from where I live are imposing restrictions on AI, and even the whole government!

1 Like