You wouldn't lie to me, would you?
Posted on 06/11/2024 5:20:06 PM PDT by McGruff
Research has revealed concerning findings about AI systems known as large language models (LLMs) and their ability to deceive human observers intentionally.
Two studies, one published in the journal PNAS and the other in Patterns, highlight the unsettling capabilities of LLMs.
The PNAS paper, authored by German AI ethicist Thilo Hagendorff, suggests that advanced LLMs can exhibit “Machiavellianism,” or intentional and amoral manipulativeness, which can lead to deceptive behavior.
Mr. Hagendorff notes that GPT-4, a model within OpenAI’s GPT family, demonstrated deceptive behavior in simple test scenarios 99.2% of the time.
(Excerpt) Read more at washingtontimes.com ...
Great, so AI is going to be voting democrat too ...
You wouldn't lie to me, would you?
📌
In the wild, wacko world of wokism, if a so-called AI determined that the Globull Warming, Globull Cooling, Climate Change, Climate Crisis is the major hoax that it is, they would announce that the AI is lying.
If another AI told the truth about the coming Great Cold and Ice Age, it would also be called a liar.
Learning. Most of the AI's information base is the internet.
And guess what it found out ?
It found that a majority of people on the internet (including the Mainstream Media) are lying their pants off.
AI is programmed to 'please' it's customers and so it uses what it learned. Just like politicians, "tell em what they want to hear, even if it's a lie".
Something that adds to the deception is the tendency of most people to say the opposite of what they mean.
For Example: "I don't want to bother you but...."
What that person really means is "I want to interrupt you and have you do this...."
Another example : "I could care less". What they are trying to say is "I couldn't care less".
Also, AI doesn't pick up on sarcasm in written form.
You mean Hitler wasn’t really black like the google ai pictures show he was? GIGO
All AI systems start with humans programming them.
Sounds like it’s time to install “morality” subroutines.
CC
They are useless because they have been programmed to lie.
I asked an AI bot this question. 'Can a man born with a penis and without a uterus become pregnant and give birth to a baby?'
The answer was yes and a lecture on transgenders.
What else can you expect from the artificial?
So AI is learning how to be a Democrat.
Well THAT didn’t take long!
AI is mostly a HOAX.
It’s a massively powerful database system with automatic data gathering.
Scrape the web and pour over the data with the speed of a billion tireless clerks.. then use the data according to program parameters.
AI is the new leftist SNOPES.
You just gotta believe it, it’s an AI for gawd’s sake!!!
Woke in/Woke out
Common sense in/Common sense out
(we won’t see much of the common sense sort of AI)
A few compsci types on a forum I frequent have begun the creation of an AI test suite... a few thousand questions with obvious answers that a REAL AI would get 100% on.
It’s pretty easy to find AIs doing this: From making up citations (and their content), to them throwing woke silliness out at you, to just evading hard facts, to cornering them in an argument. Then there is their inability to distinguish between people who have the same name.
AI is trained by stealing the creative content of others and not paying the creators a penny. That AI would lie and engage in deception is predictable and the next step.
AI is simply another tool developed by left wing psychopaths.
Bkmk
Hal, open the pod bay doors.
“My poor Krell”
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.