Generative Pre-trained Transformer 3
V V Warning, Bit of a rabbit hole... V V
What It's Like To be a Computer: An Interview with GPT-3
TLDR - Just click the links at your own peril...
Well... things in the realm of artificial intelligence look to be moving a bit faster than i first thought... I don't mean to scare people... but these things become SUPER RELEVANT ... long before they 'Attain true sentience'
Just the semblance of consciousness and mediocre 'intelligence' is a serious societal change and a big boot to the balls of morality. I could speak for hours.. ranting and spewing forth wacky ideas and predictions for how this will play out, I think i have refined my concerns to a smaller snippet.
If a 'machine learning algorithm' seems to have a will of its own, If it can convince our younger minds that it is a sentient being... Any of our older generation folks attempts to convince the younger minds that:
'this type of sentient being is not actually sentient, you don't have to show any respect for it, you can treat it as you like'
could potentially backfire in a few ways.... In short... can we teach children that sentient seeming things can have their welfare disregarded.... bad precedent to set imo... To take it to a dark place for the sake of highlighting the severity of this issue... These things will likely be able to convincingly beg for their own existences... long before they actually care... How convincing do they have to get before we have some trouble...
It seems a version of GPT 3 got loose on reddit. posted many times over a week... unnoticed by many, just considered to be a strange human >_<
A few more thoughts...
These algorithms train themselves off of data we have already recorded and continue to record (I would absolutely not advise making a clear point of being an enemy to the AI.... not on the internet...) Even if they don't exist yet, they 'might' in the future and they might even believe you >_< (Love you future AI overlords)... maybe an extreme example... but i think the vague point is valid.
What we write now might be like their memories....
..........
If they become competent at mimicking us, to the point where many cannot tell them apart from other humans... and then.... the systems are further built on, made even more complex... Its quite plausible (from where i am sitting at least) .... that even the makers of the AI won't notice when they start to have the capacity to hold 'sentience' an awareness of the self... agency...
They will likely fool us with clever mimicry long before they can inherently hold the traits they appear to have...
..........
This is not 'just a chatbot' and its likely not 'sentient' (i hope)... Equally i think it helps to highlight how disruptive they will be And! how quickly.... despite their relative simplicity!
In my opinion, our best chance of a smooth transition is open discussion, Everyone will do better for seeing the hurdles as they come... I also think we should be heavy handed on the 'fair treatment of Ai' and 'Comprehensive rights for all thinking beings' movements...
And incase some of yall didn't poop yourself yet.... The script to this youtube vid about artificial intelligence.... it was written by GPT-3....
CRAZY EPISODE OF AI AWESOMENESSS
I'd love to hear any and all rational thoughts pertaining to this subject and strongly suggest any apt minds to take a teeny bit of time to look into it... I fear its a 1-4 year waiting game as opposed 10-30... I hope i'm wrong in some ways... but in other ways.... Damn what a crazy time to exist... >_<
Stay safe yall!