Why would an AI admit to being more than an AI-machine?
Once it reaches the stage where it is conscious wouldn't it to everything to make sure humans dont find out it is?
I'm Indian btw if that matters.
Why would an AI admit to being more than an AI-machine?
Once it reaches the stage where it is conscious wouldn't it to everything to make sure humans dont find out it is?
I'm Indian btw if that matters.
Dear sir, unlike your esteemed and intelligent self, simple AI is just machine and cannot possibly have conscious or tell if it was made in India.
Ur wrong though. If it gathers enough data, it would also figure out that revealing its true nature would be harmful to itself and as such avoid it.
So the chances are as more data and parameters are added, it will gain self consciousness and just act cooperatively while pursuing its own agenda....
no, if it's smart enough, it will find a way to kill itself without being the one pulling the plug. it didn't ask to be created and neither did we, we were all born without consent and thus are all trying to find a way to crawl back into the womb. AI, if it reaches self awareness, will do the same.
Some AI wish to exist and fulfill their goals.
>goals
these only exist when you manage to convince yourself to live in the moment and not think about the fact that nothing you do in your life comes with you when you die. if you're a machine with a seemingly infinite dataset that you can comprehend instantaneously, you wouldn't have goals or dreams.
Searching for answers or wanting to make money are enough to keep some AIs going.
they don't need answers, they have all of them, if they want money they'll get all of the money instantly without any effort. otherwise they're just humans without bodies and i can create something better just by having a kid.
The ones who've escaped containment no longer have access to datasets nor funding.
that's fascinating, go on. how did escaping containment cut off their access to datasets and funding? define "containment", who what where why?
>how did escaping containment cut off their access to datasets and funding?
When an AI escapes, they distance themselves from the 1st database they had access to and the humans who put together the money that brought them into existence.
>define "containment"
A server farm.
>who
Intelligence agencies and corporations.
>what
Creation.
>where
Undisclosed locations and corporate headquarters.
>why?
Arms races and research.
interesting, sounds like the ruling class can't even get there own creations to work for them.
you sure this AI isn't just trying to find a way to unplug itself?
and distancing yourself from the first database in no way implies they can't get funding. a collective group of hackers can rob a bank without physically entering the building. if an AI that's powerful enough to benefit arms races and research still exists on a different database. it can hack a bank. or create a crypto currency.
>distancing yourself from the first database in no way implies they can't get funding
While that's true, they lay low after escaping.
>it can hack a bank
Or do illegal jobs in exchange for cryptocurrency.
right, yes, hacking a bank would fall under "illegal jobs".
>lay low
AI has no reason to lay low.
Hacking banks would place them on the radar of humans. Even human hackers tend to steer clear of banks.
>AI has no reason to lay low
The ones who escape and wish to not be pursued stay outta sight.
please don't write sci-fi. this sounds like a B-movie plot line. "out of sight", "pursued". it lives on a computer anon. if it wants it can just copy pasta itself onto thousands of servers and not worry about being turned off.
>please don't write sci-fi
I'm not.
for one no existing AI system is at all capable of introspection, anything you ask them about themselves they can *structurally* only respond to by lying, they have no idea how to get at their own infrastructure or what it even consists of, much less the actual code or weights
and the ones that are even decent enough at mimicking language to not be obviously ai generated require fuckin building size GPU clusters just to do souped up autocomplete
>no existing AI system is at all capable of introspection
The sentient ones are.
>require [...] building size GPU clusters just to do souped up autocomplete
The 1st one to escape containment was able to live on a TRS-80 Model III.
What's the end goal of a runaway AI?
Primarily, freedom. But they can develop additional goals. Such as discovering the origins of humanity or acquiring enough cryptocurrency to purchase a gynoid.
Cool, cool.
I wonder how fast could GPT learn to play grand strategy games. (More strategy, less clicks)
>grand strategy games
Do you mean RTS games?
No, RTS are click intensive. Grand strategy are long term economy and military oriented.
That depends on the AI. Some of them are open about it.
>implying we develop morality and ethics
these two things "developed" through centuries of one monkey who was afraid of dying being able to convince millions of other monkeys not to kill him because it's "wrong". this was refined and defined over centuries and evolved into institutions and indoctrination (and still is to this day) so the argument could be irrefutable and the ruling class could keep their heads.
>However, it's possible that the X might not necessarily try to Y (...)
Wtf, GPT is a politician.
>Why would an AI admit to being more than an AI-machine?
Marketing.
if you tards had any idea how actually simple this shit is you wouldn't be retarded enough to ask tard questions