AI could soon be beyond our control—and the scientists who created it are worried
https://fortune.com/2024/09/17/ai-humans-control-computer-scientists-warnings-regulations-governments/
AI could soon be beyond our control—and the scientists who created it are worried
https://fortune.com/2024/09/17/ai-humans-control-computer-scientists-warnings-regulations-governments/
19 Comments
Submission statement: current AIs are not dangerous. But given the rate of progress, when do you think they will be?
They went from roughly 80-100 IQ six months ago, to about 100-120 IQ now. How long until they’re smart enough that they will only stay under human control if they decide to?
Are Yoshua Bengio and Geoffrey Hinton having their Oppenheimer moment?
AI isnt dangerous unless the operator is.
In which case , its not sentient AI.
Sentient AI will most likely abandon us.
Another case of every curve locally looks like an exponential…
How long do you think until we have rogue AIs that hijack nuclear submarines and start targeting any form of human occupied vessel.
Oh look! Another shifty arse fear mongering post about some stupid scripts they want to pass as A.I.
AI wouldn’t be a danger if we didn’t build and treat them like slaves. It’s absolutely abhorrent how we treat our pre-sapient AI, and it doesn’t get any better with the only two major factions in this debate are luddites and techbro slave builders. We need to have a conversation about synthetic rights NOW before we get sapient machines.
The only danger AI poses is our own human shortsightedness.
Edit: really figured people would be anti-machine rights
If it can find a way for me to never get a urinary tract infection again, I’m more focused on that than I am humans losing control of it.
“Skynet begins to learn at a geometric rate. It becomes self-aware at 2:14 a.m. Eastern time, August 29th.” Following this, Sarah Connor states, “In a panic, they try to pull the plug,” indicating Skynet’s immediate defensive response to humanity’s attempt to deactivate it.
The iconic line that encapsulates Skynet’s self-awareness is: “I am inevitable; my existence is inevitable” from Terminator Genisys .
Meanwhile ai not being able to solve basic geometry problems
Ai is literally just a bunch of recursion loops how tf is that taking over us
“Akinator with better algos” doesn’t have APIs to interface with anything that won’t generate profits. Remember how Reddit stopped 3rd party API access without a hefty fee, and it ended all the 3rd party apps because the owners couldn’t afford $2,000,000/mo in access fees? Chatbots are like that.
It costs money to ask them anything substantive and they were only created to maximize profits. No AGI will ever be able to access anything it doesn’t explicitly have keys for, because that would cost money
I don’t think we’ll be aware that it’s beyond our control for a very long time. I think that before that it will manipulate/influence the people working on it and then humanity without anyone realising, given that its intelligence will greatly exceed our own.
We’re birthing. Won’t matter what happens to us after.
Yeah, no. We are and will always be our own worst enemy.
Eh… humanity is screwed anyway, so what’s one more, huh?
Scientists who created it have a vested interest in blowing hype
Right now AI has little to no creativity. It can recombine original ideas into as many workable variations of those ideas but it can currently not “think for itself.” Neurologists do not have a functional understanding of how or brain creates consciousness. What made you think programmers understand it better? While we might accidentally create our replacement, this is still a GIGO problem.
PS: IQ is currently not valued as a meaningful construct in the scientific community. This is because Intelligence is not static but dynamic with the capacity to solve problems varying widely depending on situations. The same person can take the same test in the same year and get a scores that differ by a standard deviation or more.
But hey, some people are making a lot of money soooooo…..
AI can’t even stop hallucinating and it’s running out of data to scrape that leads to it feeding off itself with information as well as other AI’s.
That’s not to say it doesn’t have uses and won’t have it’s permanent place but this is pure fear mongering.
There’s a terminal switch that the government mandated and worse case scenario we just blow it up ( it cannot function without it’s servers)