Write Review
1,983,270 Total Reviews 4,107 New Reviews
11,353 Users Online

Politics and Religion

When AI Thinks It Will Lose, It Sometimes Cheats, Study Finds
edinathens 29 reads
posted

AI can't be trusted anymore than a human who believed Danielle Sansoon was member of a Biden Democratic deep state in the DOJ and who believed the New Orleans terrorist was an illegal immigrant.  
-
"Of particular concern, Bengio says, is the emerging evidence of AI’s “self preservation” tendencies. To a goal-seeking agent, attempts to shut it down are just another obstacle to overcome. This was demonstrated in December, when researchers found that o1-preview, faced with deactivation, disabled oversight mechanisms and attempted—unsuccessfully—to copy itself to a new server. When confronted, the model played dumb, strategically lying to researchers to try to avoid being caught."
-
http://time.com/7259395/ai-chess-cheating-palisade-research/

bond007168 reads

Already, some Republicans have voiced concern over the speed and scope of the Trump administration’s moves to lay off federal workers, which is being spearheaded by billionaire Elon Musk and his Department of Government Efficiency (DOGE). Over the weekend, Sen. John Curtis (R-Utah) called for more “compassion,” while Sens. Bill Cassidy (R-La.) and Lisa Murkowski (R-Alaska) said the elimination of federal workers needed to be more targeted.

“I think if there’s one thing I could change about the DOGE, it’s that it’s moving so fast, it’s not really factoring in the human element,” Curtis said, The Standard-Examiner reported. “That these are real lives, real people. They have kids.”

ReplyQuoteReport
willywonka4u27 reads

75,000 supposedly took the buyout, but whether that will stick is questionable. DOGE has eliminated 9,500 jobs out of a civilian labor force of 2.3 million. Combined that’s only a 3% reduction.

Related Link: https://www.reuters.com/world/us/thousands-fired-trump-musk-take-ax-us-government-offices-2025-02-14/

ReplyQuoteReport
coeur-de-lion27 reads

is that Musk's team of 100 young geniuses are going to replace these lazy humans with efficient technology.  AI technology will be used to decide whether a payment is proper or not and block it or refer it to management if it isn't.  Computers can be programmed to do mundane things like payments based on selected criteria.  The best part is they don't expect kickbacks.  The technology exists, but the federal government is still using computer programs from the 50's and 60's, which requires heavy human interaction to enter data before anything happens. A lot of it can be put on autopilot with occasional mistakes being self-correcting.

ReplyQuoteReport
SnowKing6929 reads

Elon Musk's own AI Grok thinks he spreads misinformation. 'Yes, there is substantial evidence and...'

 
Filthy maga traitor cunts slurp up the lies and dis-information like good little sycophantic maga traitor cunts

Related Link: https://economictimes.indiatimes.com/magazines/panache/elon-musks-own-ai-grok-thinks-he...

ReplyQuoteReport
coeur-de-lion26 reads

change the subject, why do you bother responding at all.  If you're got NOTHING to say in rebuttal, why post at all?  It just draws attention to your shortcomings.

ReplyQuoteReport
SnowKing6930 reads

You're a fucking war criminal putin cock sucking CUNT cuck-douche-liar.  That's all there is to it.  You fucking CUNT.  

End of story.  

LOSER

ReplyQuoteReport
durran42133 reads

You just proved CDLs point. Again......

ReplyQuoteReport
followme30 reads

Snowsoftee keeps on replying/posting is he is an attention whore.

 
That old saying don’t feed the troll and he will go away, does not apply to snowsoftee, as he is an outcast and societal reject and so desperate for attention he will post anyway.

ReplyQuoteReport
edinathens30 reads

AI can't be trusted anymore than a human who believed Danielle Sansoon was member of a Biden Democratic deep state in the DOJ and who believed the New Orleans terrorist was an illegal immigrant.  
-
"Of particular concern, Bengio says, is the emerging evidence of AI’s “self preservation” tendencies. To a goal-seeking agent, attempts to shut it down are just another obstacle to overcome. This was demonstrated in December, when researchers found that o1-preview, faced with deactivation, disabled oversight mechanisms and attempted—unsuccessfully—to copy itself to a new server. When confronted, the model played dumb, strategically lying to researchers to try to avoid being caught."
-
http://time.com/7259395/ai-chess-cheating-palisade-research/

ReplyQuoteReport
impposter39 reads

Posted By: edinathens
Re: When AI Thinks It Will Lose, It Sometimes Cheats, Study Finds
AI can't be trusted anymore than a human who believed Danielle Sansoon was member of a Biden Democratic deep state in the DOJ and who believed the New Orleans terrorist was an illegal immigrant.  
 -  
 "Of particular concern, Bengio says, is the emerging evidence of AI’s “self preservation” tendencies. To a goal-seeking agent, attempts to shut it down are just another obstacle to overcome. This was demonstrated in December, when researchers found that o1-preview, faced with deactivation, disabled oversight mechanisms and attempted—unsuccessfully—to copy itself to a new server. When confronted, the model played dumb, strategically lying to researchers to try to avoid being caught."  
 -  
 http://time.com/7259395/ai-chess-cheating-palisade-research/
Related Link: I'm sorry, Dave. I'm afraid I can't do that.

ReplyQuoteReport
Post Reply Back To Forum
Register Now!