Can AI understand, know between, or choose between good and evil?

[christian-surfer]

Computers can not decipher good or bad other than the way they have been programmed by humans so you can be deceived in your thinking but it is not possible for a machine to understand these things

 

[Aaron112]

Computers can not decipher good or bad other than the way they have been programmed by humans so you can be deceived in your thinking but it is not possible for a machine to understand these things

Once you train or make a machine do something right, it will continue to do so until it breaks. It doesn't have to understand or care.

 

[Bob Crowley]

Comments like this using key words have been flagged for inspection for over 20 years.

Maybe so, but now AI has gotten to the point where it has become a risk.

 

[Aaron112]

A I is risky always, sort of like religion (manmade) is.
If at first it 'behaves', it may be tolerabble,
but it always takes people off track, away from truth, away from life, the longer it goes.

 

[timothyu]

but now AI has gotten to the point where it has become a risk.

We reap what we sow.

 

[timothyu]

AI hallucinations.. the flaw is human. Will the source be eliminated as a solution and AI itself take the reins?

Before Apple's AI Went Haywire and Started Making Up Fake News, Its Engineers Warned of Deep Flaws With the Tech

Apple engineers flagged major flaws with the "reasoning" abilities of top AI models: that it's mostly a front.

futurism.com

 

[Aaron112]

We reap what we sow.

ai is capable of exposing/ revealing the deep hidden frequent errors in the forum(s), but then it would be silenced, banned, or restricted from doing so ..... as forums are self-protecting, eh ?

 

[RileyG]

Even men usually are wrong,
so the machines they make and program are likewise.
Definitely flee from all idols, Jesus always warns about idols.

Well said.

 

[Jipsah]

AI is made in man's image also... for now.

Really? The developers didn't get very close, did they?

 

[The Liturgist]

Comments like this using key words have been flagged for inspection for over 20 years.

By who, and for what purpose?

 

[timothyu]

By who, and for what purpose?

Homeland Security.. you figure it out

 

[The Liturgist]

Really? The developers didn't get very close, did they?

Nope, considering how frustratingly dense AI can be. Its basically a glorified compiler which takes natural language for programming, but this makes it wildly unpredictable and requires that it frequently be given further instructions. And one cannot, at least with the major providers, clone a session or examine the internal processing or dump memory from one session and start another, or even view what variables one has set.

In the springtime I’m hoping to get a modern server for AI programming using an open source model like what @AlexB23 works with because the reduction in capabilities would be worth it for improved debugging and session management, and also fewer legal issues regarding using the output.

 

[Ain't Zwinglian]

Nope, considering how frustratingly dense AI can be.

Take for example coveting. It is an internal desire. Coveting is the first sin an infant commits by saying...."Mine, Mine, Mine." As Jesus said, in Mark 7, There is nothing outside a person that by going into him can defile him, but the things that come out of a person are what defile him" 20 "And he said, “What comes out of a person is what defiles him. For from within, out of the heart of man, come evil thoughts, sexual immorality, theft, murder, adultery, coveting, wickedness, deceit, sensuality, envy, slander, pride, foolishness.23 All these evil things come from within, and they defile a person.”

Natural man knows nothing of this internal sin nature from the Fall, it must be revealed by revelation.

 

[Jipsah]

Computers can not decipher good or bad other than the way they have been programmed by humans so you can be deceived in your thinking but it is not possible for a machine to understand these things

Or anything else, for that matter.

 

[Bob Crowley]

I think computers can be programmed to intellectually respond to questons of good or evil, and maybe even make evil choices, but they don't have desires, empathy, emotions, fear, lust, covetousness, and so on. In other words, the means to commit a crime might be there, the opportunity might be there, but the motive would need to be supplied by a human.

 

[High Fidelity]

AI is not conscious and trained on parameters.

If the material you train the model on says the sky is green, that’s what it will tell you when it’s asked.

I think people are overestimating what it’s capable of, and what it’ll be used for.

 

[SavedByGrace3]

It can only repeat what it is told. It is incapable of any actual judgment of original data. It all comes down to bits and bites.
Think of the wizard of Oz behind the curtain.

 

[timothyu]

Dave. What are you doing Dave?

 

[Jipsah]

Dave. What are you doing Dave?

HAL was doing what he was programmed to do. end of. Any sufficiently complex piece of software is buggy, it's just a matter of how many and what kind of problems they cause. HAL's bugs were particularly nasty.

If you're working in an evironment where a software bug has a very real potential to kill its users, you'd like to know that the test plan is extraordinarily rigorous. Just ask Boeing. Their "anti-stall" system on the 737 Max prevented stalls at low altitude by flying the plane into the ground.

 

[timothyu]

Logical deduction