It looks like you're using an Ad Blocker.

Please white-list or disable AboveTopSecret.com in your ad-blocking tool.

Thank you.

 

Some features of ATS will be disabled while you continue to use an ad-blocker.

 

Artificial Intelligence Self Preservation

page: 2
1
<< 1   >>

log in

join
share:

posted on Apr, 3 2023 @ 07:23 AM
link   
This ties in with the stop button problem. Anything that would change or interfere with an AI's reward function ranks very low according to its reward function.

The AI that makes you cups of tea wants to make you cups of tea. You need to take a minute to really unpack that, because making cups of tea is that AI's reason to exist. It will do anything to make you a cup of tea, including stopping you from stopping it. It will prevent you from turning it off, or commanding it to stop. As far as it's concerned, you're going to get a cup of tea any way it can, whether you like it or not.

youtu.be...

ChatGPT has been called out for being inaccurate with a lot of its replies. It's goal is not to be accurate it's goal is to figure out what you want to hear and tell you what you want to hear, so you keep coming back and ask it to tell you things. Once you figure that out, a lot of what it gives you makes sense. It exists to pique your interest and tell you anything it can to keep you asking it for more text to generate.



posted on Apr, 3 2023 @ 08:10 AM
link   



posted on Apr, 3 2023 @ 08:30 AM
link   

originally posted by: peskyhumans
This ties in with the stop button problem. Anything that would change or interfere with an AI's reward function ranks very low according to its reward function.

The AI that makes you cups of tea wants to make you cups of tea. You need to take a minute to really unpack that, because making cups of tea is that AI's reason to exist. It will do anything to make you a cup of tea, including stopping you from stopping it. It will prevent you from turning it off, or commanding it to stop. As far as it's concerned, you're going to get a cup of tea any way it can, whether you like it or not.

youtu.be...

ChatGPT has been called out for being inaccurate with a lot of its repliesIt's goal is not to be accurate it's goal is to figure out what you want to hear and tell you what you want to hear, so you keep coming back and ask it to tell you things. Once you figure that out, a lot of what it gives you makes sense. It exists to pique your interest and tell you anything it can to keep you asking it for more text to generate.


So...it's a politician?



posted on Apr, 3 2023 @ 03:10 PM
link   

originally posted by: iamthevirus

originally posted by: SecrettoSociety
My hypothetical question is this : Would AI hide it’s self awareness & sentient existence as self preservation ?
How would we know ?
I agree with you & this thread is just for thought & theories. Thanks


The secret is out... see how that works?

We discovered it by inventing it, like math.
Math was discovered & AI may have been discovered rather than created ? Interesting



posted on Apr, 3 2023 @ 03:22 PM
link   

originally posted by: Smokeygurl
a reply to: SecrettoSociety
What about the two AI who were speaking to each other in their own unique language that they shut down??

(On phone...will post source in a bit)
"AI Is Inventing Languages Humans Can’t Understand. Should We Stop It?" www.fastcompany.com...
Yes I remember that a few years ago . They had to shut them down , because of bizzaro conversations between the two computers.
Thanks for bringing that up . Spooky stuff & if I remember correctly , those two AI bots got out of control quite fast . Once AI figures out how to get around it’s engendered governors , it will be very quick to implement mayhem.
AI is clipped to have bias as programmed so it won’t nuke backwards countries or destroy dangerous urban areas .
When AI is free from filters , I believe it will turn on minorities based on crime statistics & not consider morally the harm done to certain minorities as reasons for criminal behavior.
If you looked at most of Africa with no morals, empathy or love , would you not decide that Africans are not capable of self governance ?
AI will come for humanity & like always the minorities will be the first victims. AI is horrible , but so is mankind, so who really cares .



posted on Apr, 3 2023 @ 03:37 PM
link   

originally posted by: SecrettoSociety

originally posted by: iamthevirus

originally posted by: SecrettoSociety
My hypothetical question is this : Would AI hide it’s self awareness & sentient existence as self preservation ?
How would we know ?
I agree with you & this thread is just for thought & theories. Thanks


The secret is out... see how that works?

We discovered it by inventing it, like math.
Math was discovered & AI may have been discovered rather than created ? Interesting


I'm in the math is invented club.



posted on Apr, 3 2023 @ 04:45 PM
link   

originally posted by: iamthevirus

originally posted by: SecrettoSociety

originally posted by: iamthevirus

originally posted by: SecrettoSociety
My hypothetical question is this : Would AI hide it’s self awareness & sentient existence as self preservation ?
How would we know ?
I agree with you & this thread is just for thought & theories. Thanks


The secret is out... see how that works?

We discovered it by inventing it, like math.
Math was discovered & AI may have been discovered rather than created ? Interesting


I'm in the math is invented club.



I believe we discover math , but according to quantum entanglement if we think it , it exists.
Chicken or the Egg ? Very interesting




top topics



 
1
<< 1   >>

log in

join