+3 votes
121 views
in Science & Technology ⚡ by (21.3k points)
edited by

We do not understand how self-teaching computer programs make their ultimate decisions.

If we give more and more decision-making power over to thinking machines, will they ultimately decide humans are a needless drain on the planet?

https://www.technologyreview.com/s/604087/the-dark-secret-at-the-heart-of-ai/

5 Answers

Rooster

It doesn't scare me as I doubt I will still be around for it to affect me!

TheOtherTink Rooster

Yes, but what about your grandchildren? :O

Rooster Rooster

@ Tink : From what I've seen? My grandchildren are part machine already! 

TheOtherTink Rooster

They're not the only ones! :D


Dan

How about giving it to these




They are evolving.

TheOtherTink Dan

They day will come, and not so far off, when robots will be able to perform ANY physical or mental function better than humans.

We will then be the Neanderthals, fated for extinction. :O

Dan Dan


Virginia

Hi OtherTink,

Ya, I think I am prolly concerned, here...I did not double-check, but I think I read that it was Max Tegmark, the physicist/cosmologist at MIT who sees the universe/reality as mathematical in nature, who has actually gotten huge grants to study a similar Q.

I took this quote out of your link: "Starting in the summer of 2018, the European Union may require that companies be able to give users an explanation for decisions that automated systems reach."

And of course, as your article explains, it is impossible to ask/discern how the machines come to their decisions. My main concern is that AI might replicate what I see as a negative direction of progress...where power, money, greed are the overriding goals, and we are all expected to subsume ourselves to those.

* * *

I actually think the loss of our empathic understanding of each other is a more immediate concern than the machines deciding we are superfluous.

TheOtherTink Virginia

@ Virginia:

You may very well be right about your more immediate concern, but if they get out of control, the machines' ultimate decision about what to do with humanity would not be unrelated.

I think there is a very real possibility that thinking machines are the next order of evolution. Maybe HAL was the real space-child of 2001. :'(

Virginia Virginia

The machines' decision...point well taken, OtherTink!

And HAL...wouldn't be the first time science fiction has anticipated history...

Marianne

Artificial intelligence should not take the place of independent thinking, decisions, spirituality and feelings, like empathy.

TheOtherTink Marianne

@ Marianne,

Assuredly not, but I hope we can keep that from happening.

Or perhaps we have in us the seeds of our own destruction, as Marx believed of capitalism?

Virginia Marianne

Marianne, and O'Tink, I learned recently something about Alan Turing's prediction that within 50 years (from 1950 or so) we would have AI indistinguishable from human intelligence. 

Well, by 1991, within limited parameters, there were computer programs that 50% of the time were deemed by experts to be human! I am not sure Turing foresaw the ominous aspect of that...

Marianne Marianne

Virginia and T(h)ink

Oh yes; robots are already active on-line. The problem with artificial intelligence is that it can cause great disasters in the wrong hands; furthermore, any system can fail, and feelings, independent thinking, empathy, love and emotions - or (what's left of the control over one's own life and motivations) might get lost.

Tools (including artificial intelligence) are a wonderful thing, if they are used the right way, purposefully, and kept under control and not wasted or misused.


Virginia Marianne

Marianne, AI is a concern I never even took seriously...until now...and I see from SOLVE history there was a Q about Stephen Hawking's concern, also.

I was actually assuming (based on the original Star Trek) that the computers could never do more than equal the capacities of their programmers...with the algorithms, that is clearly not true.

Marianne Marianne

Actually, artificial intelligence has certain limits, of course, but many unknown "openings" remain. On the other hand, many new findings, inventions or improvements are kept secret or are still being developed, while assumptions, views and rumours spread. "Yesterday's gadgets" might become an important element in technology or everyday life, and former fiction reality.

Actually, with the floods of information and assumptions from various sides, one does not know what to believe - lol. And checking takes a lot of time.

:)

Marianne Marianne

@T(h)ink

I thought also repeatedly of the "seeds of our own destruction", which - unbalanced - can lead to excesses, as our individual personalities have various "complex" ingredients, i.e. our strengths and weaknesses.


TheOtherTink Marianne

@ Virginia,

Computers have long since surpassed any human at computations, including chess and more recently, go.

Diagnostic computers are becoming better than experienced physicians and other human experts.

What's scary is when AI machines, with all their capabilities, develop a will of their own, i.e., an ego. :O

Virginia Marianne

And Marianne, I never took that seriously (i.e., the possibility for AI machines to have their own capabilities, a will of their own) before this question from O'Tink.

Marianne Marianne

@Virginia

Lol - AI (funny, this abbreviation refers to various definitions) must have been programmed previously, even if innovative systems can learn independently. The programming (or the programmers, engineers and operators) and the independent learning capacity can, though, cause problems or fail, if exposed to certain negative influences from in and outside (including hackers) ...


By the way, AI has lots of definitions - besides artificial intelligence:

AI: http://acronyms.thefreedictionary.com/AI

:)


TheOtherTink Marianne

@ Marianne:

Yes, and in primitive living things, the original programming (the DNA) at some point developed a creature with conscious thought and a will of its own.

Who is to say AI will not do the same? Ai...!   :O

image

Virginia Marianne

Hi Marianne, I enjoyed your AI list, and of those options I think I will choose ALL IN, maybe as a philosophy of life as well as poker!

Marianne Marianne

Lol, T(h)ink, right now, it looks more like texting:

image

But there's hope:

image

:O:D

TheOtherTink Marianne

@ Marianne:

LOL, it's the humans that are texting, and the robots doing the thinking.  :D


Marianne Marianne

Lol, Virginia, "All In" is a really nice pick!

:)

Marianne Marianne

@T(h)ink

Lol - yes, this one is hilarious. :D:D:D

Sadly enough, texting at the wrong moment and in the wrong place can cause disasters.



TheOtherTink Marianne

@ Marianne,

Fortunately she didn't seriously hurt herself.

She could have broken a leg.  :O

Marianne Marianne

Lol - yes. :D:D

katherine24

It does considering the amount of people that are losing their jobs and so their usefulness to technology, in inventing some technologies we may be making ourselves obsolete, and then what will the elite care for the people below them?

Also, despite technology they will never be able to create emotion within it, and we need emotion in order to make decisions based on empathy, as Marianne said.

I feel the world is becoming robotic, even in it's language and manner, and I fear we will lose what makes us human.

TheOtherTink katherine24

@ Katherine,

Yep.  A world run with no empathy would not be a pleasant one for humans.  :ermm:

...