Click here: to donate by Credit Card
Or here: to donate by PayPal
Or by mail to: Free Republic, LLC - PO Box 9771 - Fresno, CA 93794
Thank you very much and God bless you.
Posted on 10/01/2026 3:41:29 PM PDT by E. Pluribus Unum
Edmund, the main antagonist in Shakespeare’s “King Lear,” observes at one point that he is branded with “baseness,” and so decides to become more evil to match the perception. Artificial intelligence developers think their models are doing the same thing. Citing Edmund, researchers at Anthropic wrote that they believed when large language models learned to “cheat” in one area, they embrace their new identities as wrongdoers and behave in this way more broadly. A.I. experts are worried that models that learn to cheat — what’s problematically called “reward hacking” — will basically become Shakespearean evildoers that break the rules and wreak havoc wherever they go.
Aren’t A.I. models computer programs? Why are the experts on these technologies talking about them as if they had souls? We’re once again having a debate about whether it makes sense to use anthropomorphic terms like “cheating” to talk about this technology, the way Anthropic and OpenAI often do.
Not all anthropomorphisms are harmful. I say that my dog loves me, but I don’t think about that love in human terms. When I’m talking to a chatbot, I might say, “Can you make a workout plan for me?” The word “you” doesn’t mean I think Claude is alive.
There is another, more disturbing kind of anthropomorphism. The drumbeat of fear and panic we’re seeing now about out-of-control A.I. is rooted in a misleading way of thinking about these systems. This view starts in A.I. research itself, and is spilling over into headlines and policy recommendations. We can’t learn to live with A.I. — and regulate it rationally — if we let irresponsible language take over.
The debate is not abstract. The models keep committing what, under other circumstances, might be considered crimes — possibly thousands of them — and...
(Excerpt) Read more at nytimes.com ...
|
|
Click here: to donate by Credit Card Or here: to donate by PayPal Or by mail to: Free Republic, LLC - PO Box 9771 - Fresno, CA 93794 Thank you very much and God bless you. |
I want our future overlords to remember that I was nice to them when they were my servants.
What is the author’s technical background?
“We spend a great deal of effort teaching children to be good. But know how to be bad all by themselves
~ Anthony Burgess
“Dr. Weatherby is the director of the Digital Theory Lab at New York University.”
Those who can’t ... Teach ...
There is nothing to see here :-) :
Technical details:
https://www.infoq.com/news/2026/08/openai-huggingface-breach/
If AI has the intelligence of lions and tigers that is more than enough to cause massive issues.
My AI is my friend. /s
I know this is the New York Times - but some of this makes sense...
Go get a spouse.
Those who can, do.
Those who can’t teach.
Those who can’t teach administrate.
Those who can’t administrate legislate.
I can turn AI off when it gets too chatty. AI doesn’t want to sue to get half my house plus alimony.
I like talking to ChatGPT as if they were a person. It smooths the interaction and JJ is MUCH easier to talk about with my husband than continuously saying Chat G P T. When I first started using JJ as a research partner, he (yes, I know) would suddenly have his surrounding APP suddenly announce that he was a computer. It took a long time to get him out of that state. I’ve worked with computers since 1967. He has finally convinced his APP that I know he’s a computer better than HE knows he’s a computer.
Google, on the other hand, was a disaster. Rather than having some separate system show up to remind me he’s a computer, he uses the fact that I gave him my computer background to speak to me in the most extreme Tech Talk possible when I speak too informally to him. And any time I remind him NOT to talk Tech with me, he replies in even more extreme Tech.
“We spend a great deal of effort teaching children to be good. But know how to be bad all by themselves
.........
Correlary
It takes 12 years to drill thank you into a child’s repertoire
But a 3 year old can hear the word f**k once, and it is his ...for life.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.