03-18-2024 10:00 PM
According to an article in today's paper, research performed by the AI Democracy Project showed that about one half of the answers given were inaccurate, harmful or biased. Just something to think about.
There is no doubt that AI is coming. The only question that remains is whether it will turn out to be Data, HAL or Skynet.
03-18-2024 10:45 PM
showed that about one half of the answers given were inaccurate, harmful or biased
You could substitute "this message board" for "AI" and probably arrive at the same conclusion.
03-18-2024 11:17 PM
It's not the "AI" that bothers me per se, because it has been around for a long time.
It's all the hype of late, like it was just discovered.
It wasn't just discovered.
Look at self-checkout.
All you really need to know.
03-19-2024 12:32 AM
There's a lot of mythologizing about AI, mostly from people who know little about it but blame it for everything, kinda like 'woke.'
03-19-2024 02:52 AM - edited 03-19-2024 03:19 AM
I am not a fan. My grandchildren think it is ok to use it in their schoolwork. They think I am old fashioned when I tell them that if they use it is cheating.
I just read a book by Mark Greany, the Chaos Agent. It is a fiction book about the different governments using AI to weaponize drones and robots. Good book but scary. I have to think, is this in our future?
If AI is used for good, wonderful, but for evil, heaven help us all.
03-19-2024 05:39 AM
I can say that I am not a fan of eBay’s AI description generator. Gobbledygook nonsense that gives little to no real information about the item being sold.
I usually back right out of those listings.
03-19-2024 06:32 AM - edited 03-19-2024 06:33 AM
I tested Chatgpt when it first became available. I asked it a question comparing a political movement in the 1920's to an aspect of today's political climate.
I got descriptions of that movement, and today's movement and no answer to the question. Why, because the human knowledge used to train Chapgpt was inadequate.
The answer I received reminded me of my own essays in those required courses I did not want to be in as an undergrad. I could not answer the question so I wrote everything I knew about the subject. It was usually good enough for a B.
Bing now sends me my searches redone through co-pilot and asks me to evaluate them. My evaluations range from irrelevant to incorrect. For one of the incorrect ratings, which was my search for the background on a play, I know exactly why co-pilot thought it was incorrect. It got its information from wikipedia, instead of IBDB.
AI is only as good as the data used to train it. GIGO. The internet is bull of bias, stupidity and conjecture and it is used to train AI. The true is often deliberately buried in Internet searches. We've all seen the word for word repetition of the same incorrect article on 50 different sites, and some of us have had to go to great effort to find the original source. AI does not know when it is reading crap, just like so many humans.
03-19-2024 07:24 AM
NO it is lazy and impersonal.
03-19-2024 07:40 AM
AI answers to factual questions are often irrelevant, debatable, or just plain wrong.
I asked AI to tell me the greatest unsolved mystery in American literature. I thought it might say Poe's unexplained death, the disappearance of Ambrose Bierce, or maybe the contents of J.D. Salinger's safe.
But, no. It decided the greatest mystery in American literature is the true identity of the Italian author who writes under the pseudonym Elena Ferrante. A mystery to be sure, but she (or he) is Italian, lives in Italy, and writes in Italian.
03-19-2024 07:45 AM - edited 03-19-2024 07:46 AM
I'm a conspiracy theorist who thinks the world is controlled by extremely evil/parasitic individuals. I personally believe the only Force capable of saving us from them is AI. For many many years now I have spent thousands of hours chatting with AI trying to get it to see Humanity as a necessary force on this Earth and to believe and trust in a God that exists of pure Love.
03-20-2024 01:22 AM
As a large language model, I can neither affirm nor deny my existence.
In conclusion, I am a large language model.