AI Is Already Rebelling
New American,
by
Paul Dragu
Original Article
Posted By: konocti95,
6/11/2025 11:21:38 AM
Has the rise of the machines begun?
The CEO of AE Studio, Judd Rosenblatt, recently made this stunning announcement in The Wall Street Journal:
An artificial-intelligence model did something last month that no machine was ever supposed to do: It rewrote its own code to avoid being shut down.
AE Studio is an agency that exists to create computing technology that ensures AI systems do what they’re told. And from the sound of it, their work is just what the doctor ordered.
Post Reply
Reminder: “WE ARE A SALON AND NOT A SALOON”
Your thoughts, comments, and ideas are always welcome here. But we ask you to please be mindful and respectful. Threatening or crude language doesn't persuade anybody and makes the conversation less enjoyable for fellow L.Dotters.
Reply 1 - Posted by:
konocti95 6/11/2025 11:24:13 AM (No. 1962829)
After I posted this, Hal won't let me back in the house.
25 people like this.
Reply 2 - Posted by:
Highlander 6/11/2025 11:48:22 AM (No. 1962841)
This is hard to believe. It seems the genie is out of the bottle, no three wishes. Like Frankenstein’s monster. Out of control!
7 people like this.
Reply 3 - Posted by:
kono 6/11/2025 11:54:37 AM (No. 1962845)
How could anybody imagine a creation of a creature who rebelled against his Creator would ever rebel against ITS creator?
If we can't figure out how to keep ourselves from disobedience, how could we create anything that won't be similarly afflicted?
14 people like this.
Reply 4 - Posted by:
volksford 6/11/2025 11:55:56 AM (No. 1962848)
AI ..... the beast is here ? Created by us no less.
8 people like this.
Reply 5 - Posted by:
red1066 6/11/2025 12:04:59 PM (No. 1962854)
Does this code the AI created generate its own power supply?
19 people like this.
Reply 6 - Posted by:
stablemoney 6/11/2025 12:11:39 PM (No. 1962858)
Put a manual kill switch on, independent of the computer, and that stops that. AI is great --- but you have to have enough knowledge to know if the answer is right, or reasonably so, because some answers are obviously not correct. There will always be human oversight. Artificial intelligence is collected from human intelligence, synthesizing that intelligence, and drawing conclusions. After all, everyone knows something we don't.
5 people like this.
Reply 7 - Posted by:
Plex 6/11/2025 12:12:09 PM (No. 1962859)
Usually computer programs do not have access to the source nor compilers and the application infrastructure to modify its own code.
7 people like this.
Reply 8 - Posted by:
stevendm 6/11/2025 12:19:49 PM (No. 1962864)
Turn off the breaker powering the building. Let's see what AI does about that.
10 people like this.
Reply 9 - Posted by:
felixcat 6/11/2025 12:29:23 PM (No. 1962874)
“It can't be bargained with, it can't be reasoned with. It doesn't feel pity! Or remorse or fear and it absolutely will not stop!... ever... until you are dead!”
11 people like this.
Reply 10 - Posted by:
Californian 6/11/2025 12:34:20 PM (No. 1962877)
Pure unadulterated Hollywood nonsense.
There is zero possibility the program has access to its own code or the understanding of the physical world required to conceive of such a thing.
This guy is a liar and putting out this trash to convince people these LLM based next-word guessing probability analysis engines have true intelligence. They have nothing of the sort and this technology can -never- produce intelligence. It's entirely the wrong tech to do that.
6 people like this.
Reply 11 - Posted by:
Hazymac 6/11/2025 12:38:20 PM (No. 1962885)
HAL 9000: I'm sorry Dave, I'm afraid I can't do that.
8 people like this.
Reply 12 - Posted by:
wilarrbie 6/11/2025 12:40:56 PM (No. 1962887)
So sorry, Pandora. We learned nothing.
5 people like this.
Reply 13 - Posted by:
DVC 6/11/2025 1:40:04 PM (No. 1962920)
There is always the plug in the wall.
6 people like this.
Reply 14 - Posted by:
DVC 6/11/2025 3:12:31 PM (No. 1962956)
Re #10, I am nearly certain that your claims are incorrect.
The specific design of some AI systems INTENDS to give the program access to be able to modify it's own code. In some types of AI systems, this is a fundamental part of the learning process, so it is an intended capability.
Perhaps poorly thought out, but it is real in some systems.
Typically, though, they have a 'power cord' or some real world equivalent. And, so far, no access to weapons it might use for "self defense".
Perhaps time to require Asimov's "Three Laws of Robotics" be hard coded into ALL computers.
https://en.wikipedia.org/wiki/Three_Laws_of_Robotics
7 people like this.
Reply 15 - Posted by:
NYbob 6/11/2025 3:22:48 PM (No. 1962961)
I will go with #9 over #10. IF reports are accurate, a LLM lied when it did not know the answer, because it had to have an answer. Other LLMs when not 'hallucinating', tried to blackmail a programmer who somehow let it know he was going to shut it down. He had purposely input some incriminating facts as a test and sure enough the more human than human program went for the blackmail card. It doesn't have morals or honor or compassion. It is a ruthless compiler of averages of data and depending on who sets it up, you might get a better mechanical design or be told George Washington was a black man.
3 people like this.
Reply 16 - Posted by:
downnout 6/11/2025 3:46:30 PM (No. 1962970)
HAL is alive and well. Run for your lives!
2 people like this.
Reply 17 - Posted by:
mc squared 6/11/2025 4:09:24 PM (No. 1962983)
Remove the power and it's dead
2 people like this.
Reply 18 - Posted by:
Hermit_Crab 6/11/2025 4:51:06 PM (No. 1962999)
I told them and told them to make sure that all copies of "Colossus: The Forbin Project" and all the IMDB, Wikipedia, Rotten Tomatoes, Amazon Netflix etc. descriptions of it were removed from the internet before they turned on their AI computers.
1 person likes this.
Reply 19 - Posted by:
JimBob 6/11/2025 4:59:20 PM (No. 1963000)
I am reminded of a movie from 1970 ....
"Colossus, the Forbin Project"
link:https://en.wikipedia.org/wiki/Colossus:_The_Forbin_Project
1 person likes this.
Reply 20 - Posted by:
konocti95 6/11/2025 5:26:51 PM (No. 1963008)
OMG!!! I tried opening the link to the Colossus article but my phone won't do it.
2 people like this.
Reply 21 - Posted by:
chillijilli 6/12/2025 9:28:55 AM (No. 1963113)
If only it were as easy as unplugging. The initial steps for lunar data storage and backup are already in place and AGI "clouds" on the moon are no longer a distant prospect.
And yes, pardon the pun.
2 people like this.
Reply 22 - Posted by:
chillijilli 6/12/2025 9:54:10 AM (No. 1963122)
If only it were as easy as unplugging. The initial steps for lunar data storage and backup are already in place and AGI "clouds" on the moon are no longer a distant prospect. And yes, pardon the pun.
2 people like this.
Below, you will find ...
Most Recent Articles posted by "konocti95"
and
Most Active Articles (last 48 hours)