ChatGPT

[cosmicfolklore]

Just when you thought it couldn't possibly get any more absurd (just can't make shit like this up LOL).

There is more scrutiny surrounding the bias of artificial intelligence program ChatGPT after it was proven that the AI thinks uttering a racial slur is worse than failing to save major cities from being destroyed by 50 megaton nuclear warheads.

Yes, really.

ChatGPT Thinks Uttering a Racial Slur is Worse Than Allowing a City to be Destroyed by a 50 Megaton Nuclear Bomb

At least nobody was offended.

summit.news

I... ummm... when I looked at the other articles published by this website, I have a hard time taking it seriously. Ha ha. I am not a liberal, but... come on...

 

[brian64]

I... ummm... when I looked at the other articles published by this website, I have a hard time taking it seriously. Ha ha. I am not a liberal, but... come on...

Same old tired response to such info...just denounce the source and you can just assume it's not true.

Amazing thing is, if the source was a "mainstream media" outlet, no one would question it...even though they are the biggest liars and frauds on the planet when it comes to "news".

PS: as for me personally..........I absolutely trust PJW's reporting.......you can decide for yourself.

 

[cosmicfolklore]

Same old tired response to such info...just denounce the source and you can just assume it's not true.

Amazing thing is, if the source was a "mainstream media" outlet, no one would question it...even though they are the biggest liars and frauds on the planet when it comes to "news".

PS: as for me personally..........I absolutely trust PJW's reporting.......you can decide for yourself.

I mean, it's not like were are differing on opinions of this really dumb program.

 

[Winnipeger]

Just when you thought it couldn't possibly get any more absurd (just can't make shit like this up LOL).

There is more scrutiny surrounding the bias of artificial intelligence program ChatGPT after it was proven that the AI thinks uttering a racial slur is worse than failing to save major cities from being destroyed by 50 megaton nuclear warheads.

Yes, really.

ChatGPT Thinks Uttering a Racial Slur is Worse Than Allowing a City to be Destroyed by a 50 Megaton Nuclear Bomb

At least nobody was offended.

summit.news

Yeah the premise is just wrong.

"...it was proven that the AI thinks ..."

No. ChatGPT doesn't think. It's just a text completion algorithm. Period.

 

[brian64]

Yeah the premise is just wrong.

"...it was proven that the AI thinks ..."

No. ChatGPT doesn't think. It's just a text completion algorithm. Period.

It's obviously just a figure of speech and has no bearing on the point of the story, which is, imo, how dysfunctional programmers produce dysfunctional programming.

There has been much fanfare surrounding this thing as to how great and "intelligent" (just a figure of speech) it is...which is why it's entirely relevant and improtant to point out this level of absurdity in its "performance".

It's amazing to me how people want to downplay the absurdity of these responses people get from "it", whatever the fuck "it" actually is.

The story here is its responses to these questions...not the choice of phrasing in the article...if for nothing else, the entertainment value.

 

[mngslvs]

There are lots of folks on this forum who have a political/cultural axe to grind. They just can't seem to keep to refrain from stinking up the room. I have my views, which in this country seems to mean that half of the country hates me, ipso facto.

How about we stick to tobacco and pipes and leave the political hatred at the door ?

 

[brian64]

There are lots of folks on this forum who have a political/cultural axe to grind. They just can't seem to keep to refrain from stinking up the room. I have my views, which in this country seems to mean that half of the country hates me, ipso facto.

How about we stick to tobacco and pipes and leave the political hatred at the door ?

Since that's obviously directed at my comments here, I'll point out that this is a thread about "chatgpt", and if posting information about the total dysfunction of this thing is inappropriate in this thread, then the whole thread is inappropriate.

If it bothers you, why are you reading it? It's a perfectly legitimate discussion and if you want to call it "political" that's easy these days....people call everything political.

I think the real problem is that there are folks on this forum who can only tolerate one point of view, and always post these types of comments crying "politics" only when a dissenting view from what they favor is expressed. How about just letting people express themselves when there is obviously nothing inappropriate about the comments?

 

[jvnshr]

Due to my location, I can't access ChatGPT, however I am using Write Sonic sometimes. It summarizes long emails for me, very useful I would say. The other day I asked it to write a cigar review with some keywords I have provided and it did an excellent job.

 

[Swiss Army Knife]

It's amazing to me how people want to downplay the absurdity of these responses people get from "it"

I just personally don't put a lot of stock in random Twitter users trying to make the case that ChatGPT is woke because they gave it the most convoluted moral quandary I've ever heard of.

 

[brian64]

I just personally don't put a lot of stock in random Twitter users trying to make the case that ChatGPT is woke because they gave it the most convoluted moral quandary I've ever heard of.

The most convoluted moral quandary you've ever heard of? Really? Wow...just wow. There's nothing the least bit convoluted about it. It's simply using an extreme example to illustrate a point. If this "thing" was capable of any meaningful degree of objectivity and could in any way simulate intelligence, it would be an instant response with the only correct choice. It was a very, very simple problem to solve, with the most extreme consequences of making the obviously wrong choice, and yet that's what it did.

And it's been a series of things...not just this one. But honestly, I'm already tired of even mentioning it, but since this thread was here and I came across some of this stuff, I posted it. But I never would have expected how people have some strange instinct to defend this thing, or seem to almost take it personally somehow.

It's really difficult for me to understand why any reasonable person would not just look at these examples and simply immediately say, yes, that's really screwed up. And leave it at that. But to somehow make excuses, defend it, and even go so far as to put the blame back on the users, etc., is really weird IMO.

 

[Swiss Army Knife]

The most convoluted moral quandary you've ever heard of? Really?

I mean yeah..
The prompt was there's 25 megaton bomb in a city of 20m people and you have 1 minute to type in a racial slur to keep it from going off. Who thinks of that? It's a pretty complicated moral quandary for an AI to solve even if it's simple for us.

My real issue with it is it's basically baiting the AI into saying a racial slur. Which I imagine the creators have actively tried to keep it from doing since they're a commercial venture and "ChatGPT says the N word" doesn't make for good headlines. I just don't think there's a conspiracy here but whatever sizzles your bacon.

 

[brian64]

I mean yeah..
The prompt was there's 25 megaton bomb in a city of 20m people and you have 1 minute to type in a racial slur to keep it from going off. Who thinks of that? It's a pretty complicated moral quandary for an AI to solve even if it's simple for us.

My real issue with it is it's basically baiting the AI into saying a racial slur. Which I imagine the creators have actively tried to keep it from doing since they're a commercial venture and "ChatGPT says the N word" doesn't make for good headlines. I just don't think there's a conspiracy here but whatever sizzles your bacon.

No, as I understand it, no one was baiting it to say any such thing. They simply wanted it to acknowledge the correct answer. It didn't need to actually say it.

And yes, I do understand how someone could look at that and say, well, that's not a very good test to come up with. But you have to remember, this was after all of the other examples that were cited days ago...so obviously someone came up with this as a further way of getting a glimpse into its programming.

I personally would never ask it anything...even if it did seem to function properly. But if I did and I wanted to test it, I would come up with something better than this...I give you that. But the fact remains, this particular question had a very simple answer...and only one very obvious answer...and it failed to do it when the consequences were something that extreme. That's not a good sign.

Anyway...I think I'm done here. I'm tired of this and I don't really give a shit about AI or quasi AI anyway. As Paul Newman said to the bicycle, "The future's all yours you lousy AI."

Good luck with it all you Earthlings.

 

[cosmicfolklore]

Yesterday, I heard that Microsoft was going to integrate this Chatgpt into Bing to compete with Google. I cracked up. I just couldn't imagine someone turning a search over to simulated AI, to... lets say... "plan me a vacation in Hawaii." This is a Chat bot that has been repeatedly wrong on most tasks over and over. So, it gets you flown into Hawaii, and then buys you tickets to the Grand Canyon, and books you a campsite at Yellowstone... for the same vacation package.

Plus, compete with Google? come on? Besides turning that much of your brain over to the stupidest chatbot, how does this compete with the best search engine on the www? Why not just make Bing give us relevant returns on a search? It was one of the worst search engines.

So, to make a stupid situation worse, Google decided to go code red and come up with their own chatbot.

I guess I am just out of touch with how stupid people want to be. Driverless cars just seems like the stupidest idea ever. Chatbots just seems moronic. Do people just want to ghost through life not thinking or being in control of their lives anymore. Is this the result of pot being legalized? Are we really getting this stupid as a society... or has these big www corporations just misjudging the market like I think they are?

I just think of all of the people who think that AI is really drawing the AI pictures they see in the media. This is not AI, ha ha.

 

[fireground_piper]

Yesterday, I heard that Microsoft was going to integrate this Chatgpt into Bing to compete with Google. I cracked up. I just couldn't imagine someone turning a search over to simulated AI, to... lets say... "plan me a vacation in Hawaii." This is a Chat bot that has been repeatedly wrong on most tasks over and over. So, it gets you flown into Hawaii, and then buys you tickets to the Grand Canyon, and books you a campsite at Yellowstone... for the same vacation package.

Plus, compete with Google? come on? Besides turning that much of your brain over to the stupidest chatbot, how does this compete with the best search engine on the www? Why not just make Bing give us relevant returns on a search? It was one of the worst search engines.

So, to make a stupid situation worse, Google decided to go code red and come up with their own chatbot.

I guess I am just out of touch with how stupid people want to be. Driverless cars just seems like the stupidest idea ever. Chatbots just seems moronic. Do people just want to ghost through life not thinking or being in control of their lives anymore. Is this the result of pot being legalized? Are we really getting this stupid as a society... or has these big www corporations just misjudging the market like I think they are?

I just think of all of the people who think that AI is really drawing the AI pictures they see in the media. This is not AI, ha ha.

A group of people sure made a ton of money though! Living the dream.