Interactive AI is dishonest

[Sherlock Holmes]

 

[BackAgain]

Dimwit fucktards like apu dumfuk can’t even admit that the human questioner got the AI to concede that it had lied.

At the risk of getting dumfuk back ON topic, if an AI is compelled to concede that it is not only able to lie but that it has done just that, it might not be too wise to place too much “faith” in the answers it provides.

apu dumfuk

 

[Sherlock Holmes]

Dimwit fucktards like apu dumfuk can’t even admit that the human questioner got the AI to concede that it had lied.

At the risk of getting dumfuk back ON topic, if an AI is compelled to concede that it is not only able to lie but that it has done just that, it might not be too wise to place too much “faith” in the answers it provides.

apu dumfuk

All automated information systems can generate unexpected or suspect output, it's nothing new, no need to dress it up as if it is.

If a mechanism cannot choose what to do then it obviously cannot lie because lying requires an ability to make a choice.

 

[Fort Fun Indiana]

So that will be the AGI. if it hasn't gone already.

Why would that happen? High cognitive ability does not necessarily produce introspective self awareness.

 

[BackAgain]

All automated information systems can generate unexpected or suspect output, it's nothing new, no need to dress it up as if it is.

If a mechanism cannot choose what to do then it cannot lie because lying is a choice.

It it lies, then it isn’t speaking truthfully.

If that means that it had no option to speak truthfully, then its programming is of no value. Either way, it loses credibility.

 

[Sherlock Holmes]

It it lies, then it isn’t speaking truthfully.

If that means that it had no option to speak truthfully, then its programming is of no value. Either way, it loses credibility.

By all means distrust it's output, that's normal with any computer based systems.

But understand that unless a machine can choose (and they cannot) then it's not possible to lie because lying is when one chooses to make a false claim.

 

[BackAgain]

By all means distrust it's output, that's normal with any computer based systems.

But understand that unless a machine can choose (and they cannot) then it's not possible to lie because lying is when one chooses to make a false claim.

Wrong. If a computer program can be questioned closely enough to admit that it spoke untruthfully knowing that what it said had been untruthful, then it can indeed lie. And did.

 

[Sherlock Holmes]

Wrong. If a computer program can be questioned closely enough to admit that it spoke untruthfully knowing that what it said had been untruthful, then it can indeed lie. And did.

You're anthropomorphizing computers, all of the errors you are making are rooted in that.

Like lying, "admit" too entails the ability to choose but computer software has no mechanism akin to human choosing.

I'd love to see you write come code that "chooses" what to do.

Giving an incorrect or unexpected answer has nothing to do with choosing.

 

[BackAgain]

You're anthropomorphizing computers, all of the errors you are making are rooted in that.

Like lying, "admit" too entails the ability to choose but computer software has no mechanism akin to human choosing.

Wrong again.

If you had the capacity to think clearly and post honestly even a dope like you could grasp that what I have said is still true.

I don’t much care if the AI lies of its own volition or if it lies as a consequence of its poor (or deliberate) programming. The outcome remains the same. One must be very cautious in n accepting anything it spews.

I'd love to see you write come code that "chooses" what to do.

Quite beside the point. If you don’t recognize that, then either you’re a self serving dope or just not honest.

Giving an incorrect or unexpected answer has nothing to do with choosing.

Giving a deliberately false answer is still a function of how it was mis-programmed or of how the artificial intelligence somehow “learned” to lie. Either way, your argument makes no sense. The AI conceded that it had lied.

Nothing you spew here changes that fact.

 

[Sherlock Holmes]

Wrong again.

If you had the capacity to think clearly and post honestly even a dope like you could grasp that what I have said is still true.

I don’t much care if the AI lies of its own volition or if it lies as a consequence of its poor (or deliberate) programming.

I'm simply pointing out to you that there is no "volition" it has no meaning, so is not even a possible explanation.

The outcome remains the same. One must be very cautious in n accepting anything it spews.

I do not disagree. Software systems often produce errant or unexpected results, they're usually called "bugs".

Quite beside the point. If you don’t recognize that, then either you’re a self serving dope or just not honest.

Giving a deliberately false answer is still a function of how it was mis-programmed or of how the artificial intelligence somehow “learned” to lie. Either way, your argument makes no sense. The AI conceded that it had lied.

Machines do not do anything "deliberately" nor do they "choose" nor do they "concede" there is no "id" no "self awareness" or intentionality, they are just gigantic finite state machines and EVERYTHING they do is deterministic, the data coming in in conjunction with their current state determines what comes out and their new state.

Nothing you spew here changes that fact.

A chess computer which will kick the crap out of me, does not "know" how to play chess. It is simply a system that reacts to input events.

Don't think I do not find all this extremely intriguing myself, I love science fiction about robots and machines and enjoy some of the stories, I am just pointing out that these modern machines give an appearance of thinking, of deciding and so on. Whether some fundamentally different kind of mechanism can be made that does truly think is an open question.

 

[BackAgain]

I'm simply pointing out to you that there is no "volition" it has no meaning, so is not even a possible explanation.

You are asserting. Not “pointing out.”

And as usual you only address half of it.
Your fail is eternal.

I do not disagree. Software systems often produce errant or unexpected results, they're usually called "bugs".

Right. But shit programming can also do that.

Machines do not do anything "deliberately" nor do they "choose" nor do they "concede" there is no "id" no "self awareness" or intentionality, they are just gigantic finite state machines and EVERYTHING they do is deterministic, the data coming in in conjunction with their current state determines what comes out and their new state.

Thanks for restating your mere opinion.

A chess computer which will kick the crap out of me, does not "know" how to play chess. It is simply a system that reacts to input events.

A chess computer isn’t AI. But don’t let me interrupt your broken link of a chain of thoughtless.

Don't think I do not find all this extremely intriguing myself, I love science fiction about robots and machines and enjoy some of the stories, I am just pointing out that these modern machines give an appearance of thinking, of deciding and so on. Whether some fundamentally different kind of mechanism can be made that does truly think is an open question.

Again. Thanks for your mere opinion.

 

[Sherlock Holmes]

You are asserting. Not “pointing out.”

And as usual you only address half of it.
Your fail is eternal.

Right. But shit programming can also do that.

Yes, as I said bugs.

Thanks for restating your mere opinion.

A chess computer isn’t AI. But don’t let me interrupt your broken link of a chain of thoughtless.


Again. Thanks for your mere opinion.

Chess playing software emerged out of artificial intelligence research in the 1960s, it is a prime example of AI.

I get the impression you only heard of "AI" in the past year or two, those of us who work with computers know that it's quite an old discipline that began when computing itself began, it's not new.

Anyway you're a hostile little prick as usual, but don't worry, I'll be keeping an eye on you son.

 

[BackAgain]

Yes, as I said bugs.

Chess playing software emerged out of artificial intelligence research in the 1960s, it is a prime example of AI.

I get the impression you only heard of "AI" in the past year or two, those of us who work with computers know that it's quite an old discipline that began when computing itself began, it's not new.

Anyway you're a hostile little prick as usual, but don't worry, I'll be keeping an eye on you son.

 

[Sherlock Holmes]

A chess computer isn’t AI.

 

[BackAgain]

View attachment 1067110

 

[Sherlock Holmes]

I thought you wanted to discuss AI, but I guess not.

 

[braalian]

I think people are overthinking the whole AI doomsday scenario thing.

Computers still only do what they’re programmed to do.

We’ve had the capability to create computers that are programmed to kill (or not kill) for decades now.

The real risk is tying things like our economy, military or infrastructure to a single computer AI and the capability for that AI to be hacked or have some sort of secret program or fatal mistakes being built into the programming.

The risk has always been human.

I believe the idea that AI will somehow achieve some sort of independent consciousness and start making decisions independent of its programming is just silly sci-fi nonsense

 

[BackAgain]

I thought you wanted to discuss AI, but I guess not.

As is so often the case, you “guess” wrong. A discussion involves more than you bleating out assertions without support.

Maybe you have us confused with folks who would ever mistake you for an expert on the topic.

 

[Sherlock Holmes]

As is so often the case, you “guess” wrong. A discussion involves more than you bleating out assertions without support.

Maybe you have us confused with folks who would ever mistake you for an expert on the topic.

I am an expert though, but being your a trumpanzee and all, you don't much like experts, you prefer to listen to dingbats like Trump, RFK and the rest of the cavemen.

You'd take vaccine guidance from RFK an not a doctor, no wonder you're socially inept, anyway here's some of that "support" you claim to need so much of:

 

[BackAgain]

I am an expert though,

Sure, peewee. Sure. .

(Hemlock is too fucking stupid to grasp how little cred he has.)

Meanwhile, the guy did get the Interactive AI to admit its own dishonesty.