How Much Should We Trust Artificial Intelligence?

MineSubCraftStarved

@JulesKorngold
It all depends on how artificial intelligence is programmed. It is not possible for any program or deep learning application to go beyond its boundaries and do something other than what it was programmed to do.
There is no chance for some deep-learning software that is designed let's say to write essays, to then also go and take over major social media applications since it has absolutely no ability to do so.

Nomenclature

@JulesKorngold

About as much as artificial intelligence should trust us, the single biggest threat to the survival of the planet.

jack

MineSubCraftStarved said:

It all depends on how artificial intelligence is programmed. It is not possible for any program or deep learning application to go beyond its boundaries and do something other than what it was programmed to do.

Hello M:

What makes AI intelligent, is its ability to recognize its own mistakes and write its own fixes for them.  Ergo, it'll eventually SEE us as a mistake.  And, from a computers point of view, who can argue with that?

excon

Nomenclature

@jack

What makes AI intelligent, is its ability to recognize its own mistakes and write its own fixes for them

Which makes it far more intelligent than you.

piloteer

jack said:

What makes AI intelligent, is its ability to recognize its own mistakes and write its own fixes for them.  Ergo, it'll eventually SEE us as a mistake.  And, from a computers point of view, who can argue with that?

excon

If AI is only recognizing its "own mistakes", exactly how could it consider humans a "mistake" when all its doing is recognizing its "own mistakes? It's kind of a leap to go from "fixing its own mistakes", to "seeing us as a mistake", and vaguely suggesting the only obvious outcome would be global destruction of humans.

Strict government mandates to ensure transparency of every step of every process can help ensure safety. It would be ignorant to try and bann it in the US, while other countries that are 10 years ahead of us technologically are freely experimenting with AI and keep us out of the loop. The best thing to do would be to ensure full public transparency of the process instead of trying to hide from it, which just allows others to make their own AI toys in the dark, and without any of our permission.            

MayCaesar

There is nothing special about artificial intelligence: it is just another technological milestone, in the row of hundreds of milestones that in the past were expected to bring the civilization to its knees. At its core artificial intelligence is just a system of converting sensory inputs into actions, same as any other system, just more complex than most other systems in its behavior. The idea that a rogue AI can wreak havoc on human systems, let alone conquer or destroy humanity, seems utterly preposterous to me, as well as the claim that it will cause mass unemployment, runaway wars, "singularity" and so on. An AI, at the end of the day, is just an algorithm, similar to the one running in a human brain, and in itself it has absolutely no power to accomplish anything: it has to have access to the real world physical resources to do some damage, and that access is not easily obtainable or maintained. An AI will not be able to, say, take control of the Russian nuclear arsenal and start a nuclear war, considering how many steps it involves and in how many ways one single step can go wrong making the allegedly desired outcome impossible. For that matter, the idea that an AI would even want to destroy or seriously harm humanity is completely unfounded: it seems to be about as likely as a human that would want to destroy all technology and revert humanity back to the ice age - might be some weirdos around genuinely wanting that, but they are far outcompeted by the billions of those who want nothing like that to happen.

Sorry, folks, but, as exciting as it would sound, no Skynet is coming to Earth in the foreseeable future.

JulesKorngold

Argument Topic: The Core Of Artificial Intelligence

AI is primarily achieved by reverse-engineering human capabilities and traits and applying them to machines. At its core, AI reads human behavior to develop intelligent machines. Simply put, the foundational goal of AI is to design a technology that enables computer systems to work intelligently yet independently.

JulesKorngold

Argument Topic: An Example Of Powerful And Successful Artificial Intelligence

AI has drastically improved traveling. Instead of having to rely on printed maps or directions, you can now use Google or Apple Maps on your phone and type in your destination.

So how does the application know where to go? And what’s more, the optimal route, road barriers, and traffic congestions? Not too long ago, only satellite-based GPS was available, but now, artificial intelligence is being incorporated to give users a much more enhanced experience.

Using machine learning, the algorithms remember the edges of the buildings that it has learned, which allows for better visuals on the map, and recognition and understanding of house and building numbers. The application has also been taught to understand and identify changes in traffic flow so that it can recommend a route that avoids roadblocks and congestion.

jack

piloteer said:

If AI is only recognizing its "own mistakes", exactly how could it consider humans a "mistake" when all its doing is recognizing its "own mistakes? It's kind of a leap to go from "fixing its own mistakes",

Hello again, p:

I dunno what it knows or if it knows anything at all..  But, it DOES know cancer when it sees it.  It knows our blood pressure.  It knows our hearts fail.  It knows that we get old.  I dunno if it'll eventually get that we're breakable, and it's not.  And, I dunno if it would consider that a "mistake"..

excon

Nomenclature

@jack

Hello again, p:
I dunno what it knows or if it knows anything at all.

If it has an IQ above 65 it knows you're not Jewish. 

John_C_87

@JulesKorngold

AI is primarily achieved by reverse-engineering human capabilities and traits and applying them to machines. At its core, AI reads human behavior to develop intelligent machines. Simply put, the foundational goal of AI is to design a technology that enables computer systems to work intelligently yet independently.

Just to make something clear you want computer programming to be the industry where international or accidental negligence is not criminal because a computer driven machine is independent? An example of this kind of incompetence is someone saying as an answer to a question, “You are crazy.” Or another example might be “I can understand you.” 

John_C_87

If it has an IQ above 65 it knows you're not Jewish. 

As someone who has gone on record as saying they can't tell the difference between facts such as ratio and approximation. What multiple quantities are divided in a computer to deliver a skill of reason to try, learn, understand new things. / I. Q.? 

Computers do not have a I.Q. it is 0 meaning (Zero) Dee. The problem is in the interpretation of some human laws around computer technology as there is a connection that has been made to law and machine with firearm legislation which is establishing a direct line to established justice.

John_C_87

@JulesKorngold

Overall, whether or not we can trust artificial intelligence depends on how the system is designed, trained, and implemented, and how well it aligns with ethical principles and values. It is important to evaluate AI systems carefully and ensure that they are used responsibly and ethically.

I disagree JulesKorngold, it is not about the trust of AI it is about the precedent set by a firearm which connects the programmer to the computer as a person who pulls a trigger. And. Furthermore. It is mistakes such as Pi and the grievance of natural and real numbers which set a liberty on that burden of legal precedent by sharing the weight of related death with those who instruct the programmer. You make it sound as if when a person finds a error if negligence they do not have a burden to report it upon its finding, A.I. is trusted fully because it does not exist the intelligence is the programmers...