Artificlal Intelligence

[kerouac2]

AI is something that probably most of us never expected to become an issue, but suddenly it is.  AI is writing college papers, novels, scenarios, things for which many of us would prefer that it not replace human beings (or allow them to cheat).

I have not yet made an opinion on it and I have not consulted chatGPT (yet) because the idea frightens me a bit.

At the same time, it seems to be so well programmed that it is more reasonable than most human brains, so that isn't bad, is it?  It seems a lot better than a lot of our politicians.  Then again who will be programming the AI of the future?  Right now, the United States and China dominate this technology, but what about in a couple of years if Russia and Saudi Arabia and Brazil invest in it?  Each country injects its own ideology into AI.

At least the EU is already making an attempt to make some rules.  What do you think?


How Europe is leading the world in building guardrails around AI

[onlyMark]

AI is a Pandora's Box that at some time had to be opened and now it is too late to close it.
I've used ChatGPT out of curiosity several times and the answers it gives are certainly more readable than a straight google search - but I think unless you pay for the most recent model you are restricted to the knowledge base used which is only up to 2021.
The biggest problem is that it makes things up. Can be facts and figures it makes up and/or the reference work those come from.

Making laws is useless because AI development is irresistible to many and they will find a way around them. Imagine North Korea having an advanced AI and the havoc it would cause as they certainly won't abide by any restrictions. That will mean other countries will need to develop an AI that is capable enough to overcome the 'rogue' one.
It is now an arms race and nothing can be done about it.
It has been said about being able to identify an AI using the Turing Test. But that will not work.
Remember that any AI intelligent enough to pass the Turing Test is also intelligent enough to know how to fail it.

ChatGPT and it's further iterations (up to Chat GPT-4?) knows how to code and is connected to the internet. It won't be long before it can/will seed itself to places it cannot be restricted.
Isaac Asimov in 1942 predicted there should be rules/laws written into robots and suggested, "A robot may not injure a human being or, through inaction, allow a human being to come to harm. A robot must obey orders given it by human beings except where such orders would conflict with the First Law. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law." - maybe that is a start.

As Haddaway said, "Life will never be the same, life is changing."
Also - "Private James Frazer is a fictional Home Guard platoon member and undertaker, first portrayed by John Laurie in the BBC television sitcom Dad's Army. He is noted for his catchphrases "We're doomed!"
Arthur C. Clarke also realised the dangers of an AI (HAL, 2001: A Space Odyssey).
When an AI is paired with a proper quantum computer, that is truly dangerous. If it then achieves sentience, that is our death knell.

[kerouac2]

I never forgot the British science fiction novel Colossus which I read in my youth.

The story is set in the 1990s. Computer expert Charles Forbin briefs the President of the United States of North America on the completion of Project Colossus, a powerful system designed to control the USNA's nuclear arsenal from within the Rocky Mountains. Although the President eagerly relieves himself of the burden, Forbin voices doubt about conferring absolute military power to a machine. The President announces Colossus to the world; his government expects that giving irreversible control of USNA nuclear weapons to an unemotional, logical, impregnable computer will make the country and world safer.

Of course there is a similiar computer in the Soviet Union, so here we go...

[mossie]

Let us not forget that Colossus was the name of the original computer made at Bletchley Park, by British Post Office engineers to decode the German Enigma messages during the war

[lugg]

I don't really understand much about AI yet but what I have read makes it sound quite alarming. I have not looked at ChatGPT.

I think apart from the news I mostly became aware of it when I was viewing some photos on a group I subscribe too ( underwater images) and a couple were called out as AI images. That really stumped me and made me want to learn more

[bjd]

Artificial intelligence has been around for decades, for example, used in flying planes. Or did you think the pilots are hands-on during the entire flight? But the subject was of interest to engineers and scientists, not the general public.

Now, like so much else, the term AI is thrown out into public view with no real information but with lots of scare-mongering and misinformation.

[onlyMark]

I'm not sure you class the computer in planes as AI. An AI learns, develops and adapts.

[bjd]

I met several people in Toulouse who worked for Airbus, as well as PhD students working in the computer field, and they all called computers in planes artificial intelligence. We even had a copy of a book one friend wrote on the subject back in the 1980s.

[kerouac2]

I would say that AI begins when fuzzy logic is involved, so it definitely exists with aircraft systems, but that is still an extremely primitive level of AI and not at all what is setting the media on fire at the moment.

[onlyMark]

You could be right bjd but I agree with K2.
Computers on planes use machine learning, which though is a a part of AI, isn't AI itself to me.
"The main difference between artificial intelligence and machine learning is that AI is a general system with cognitive capabilities, and machine learning is how systems sort through data sets and build intelligence and cognitive-like abilities."

Airbus say, "Computer vision and machine-learning technologies based on AI are critical to enabling self-piloted commercial aircraft to take off and land, and to navigate and detect ground obstacles autonomously."
Based on AI, using features of AI, but not true functioning AI is how I interpret that.

[bixaorellana]

If anyone is interested, there is a novel about AI by Ian McKewan on sale now in the kindle edition.
Go here: anyportinastorm.proboards.com/post/408865/thread (Reply #343).

[whatagain]

I just read this :

I am fearing less the increase of artificial intelligence than the decrease of human intelligence.

Was funny, but makes one think...