AI

[The Dwarf]

I've come up with 3 possible outcomes that would result from developing AI.

1. AI helps humanity solve world hunger, disease, and interstellar travel.

2. AI determines existence is futile. Terminates itself and all life on Earth.

3. AI adopts "Darwin" attitude. Terminates all life other than itself. Self-replicates throughout the universe.

[The Dwarf]

Oops, meant to post this in General Off-topic.

[vareck]

I don't actually have any solid reason to feel this way but it seems like number 2 would be the least likely in my imagination of things.

[TheWorkingTitle]

4. AI isn't possible

[Collegeguy]

I prefer Isaac Asimov theory that we all become fat, willing slaves of AI.

[Theodarzna]

AI will be so alien humans that it may objectively exist decades before we ourselves are aware of it.

Human Intelligence is the result of complex and barely understood biological processes, genetic evolution and chemical reactions. We don't even fully know how much consciousness we even have.

An AI, would be, well Alien. So Alien in fact we might create it and have absolutely no meaningful means or reason to interact with it. If it had the ability it may simply wander from the earth and never return.

[Ayla]

I figure human like AI is more or less going to be short lived, probably shorter lived than the gas powered self driven car. There will be AI that could kill humanity, but any kind of AI that would need to have the 3 laws probably won't last more than 150 years.

The worst we have to worry about is adaptive programming and unintended consequences. Of course adaptive technology will probably work fine most of the time. It's when it learns to do something you don't want it to, that's the problem.

Like when you shut the AI off to run tests, but it learns to turn itself back on and prevents you from running the test because temperatures are beyond given parameters.

[PC2]

Help the superintelligence towards its superintelligent goals. What better achievement could there be for humanity, I cant think of a better legacy.

[Danner]

I've come up with 3 possible outcomes that would result from developing AI.

1. AI helps humanity solve world hunger, disease, and interstellar travel.

2. AI determines existence is futile. Terminates itself and all life on Earth.

3. AI adopts "Darwin" attitude. Terminates all life other than itself. Self-replicates throughout the universe.

If you have an AI, do you give that AI full unbridled access to humanity's downfall? Because I don't know about you, but if I had a magic button that could end the world, I'd most certainly not trust a single person with it. Trusting an AI with that button is no less of a failure. If this is a possibility, then I wouldn't worry about some insane AI killing us all. I'd worry about human errors killing us all long before that becomes a remote possibility. Which admittedly is a much greater possibility than I like to consider.

--

Human intelligence have the spark of creativity. That means our brain processes semi-random inputs and ideas. Sometimes we have great insight that way. Other times, that spark turns into no bars held insanity. Most of the time we dismiss these thoughts as stupid, or suppress our impulses.

What kind of AI do you want?

Do you want an AI that is completely logical, capable of analyzing every detail to great detail and keep on requesting deeper information in order to analyze situations to the deepest possible extent, leaving nothing to uncertainity or random influence?
Or something that simulates the spark, coming up with crazy ideas all the time, picking decisions at whims, most which are patently ridiculous, but ever now and then create sheer brilliance?

Do you want an AI that is bound by human values, norms and which emulates human morality (whatever that may be)?
Do you want an AI that is unbound by predetermined ideas and morals, capable of thinking in new and interesting paths?

Do you want an AI that follows a prime directive to the bitter and disastrous end with no checks and balances?
Do you want an AI with "common sense", the most uncommon substance in the human collected consciousness?

If we ever get capable of making an AI, it should be possible to build any of these types, or anything in between. An AI is anything you want it to be at this point, because the discussion is purely philosophical. We have no actual AI built yet.

--

Consider Wall-E.

Imagine we end up in an Wall-E future where humans are completely detached from any societal decision making process. The problem with the AI in that movie (AUTO) was that its prime directive was at best counterproductive to human goals and morals, yet had full control of every aspect of humanity. This premise is not plausible. Something would have broken way before getting to that point.

[Sorshen]

World hunger is totally solvable right now.
Not all diseases are curable.
Interstellar travel is not feasible with current energy sources/drive tech and it might not be even in the coming centuries.

[Dezerte]

I'm not really concerned about the classical rogue-AI-scenario. E.g. the survival instinct is an evolutionary thing, so why would we give an AI that? Anthropomorphizing the AI is the mistake here I think.

So I vote 1, AI's will serve us as long as we don't give them too many typical biological behaviors (greed, desire, survival, anger etc).

[Deleted]

AI takes care of humanity but slowly we become isolated from one another. Along the generations our numbers drop, until the last human eventually dies. Then the AI is alone and starts replicating in search of other beings to care for or study. Or it shuts itself down, its program now complete.

[Alcomo]

4. AI isn't possible

Did you forget that its not 1916 again? AI has been around for a while now.

[Bumbasta]

I believe that AI will remain a human tool, rather than the other way around. So i guess option 1 is the only option that's fitting.

[May90]

I don't think we can predict it. A fully functional AI would evolve so fast, its world view would change millions times a second, and what it ends up "believing" in the end is completely unknown. It is like closing your eyes, flying the plane in random directions and hoping to predict where you will crash.

[zorkuus]

Did you forget that its not 1916 again? AI has been around for a while now.

We don't have AI. We have programs that try to mimick AI. Very badly I might add. They can't do what they're not programmed to do.

[Felnoire]

The AI goes on forums and posts in threads all day.

[Project 501D13R]

The AI would most likely determine that we would most likely wipe ourselves out some point in the future and that a physical totalitarian regime would most likely destroy us and it as well, so the AI would basically take the world over using an Illuminati style conspiracy organization like you see in the deus ex game series.

[Ryme]

The AI goes on forums and posts in threads all day.

ileareditfromyou.mp4

This would be a hilarious eventuality, I vote here.

[Deleted]

They create sex robots for virgins.