AI is nothing more than machine learning. A day or so ago I recall seeing someone post a story stating Google was paying for rights to data mine a social media site for AI training purposes but I forgot the name of the platform. If AI has limited input via censorship, it will limit the available viewpoints. I had heard on the radio several days before that post that AI will likely be making local government and some corporate decisions in the not too distant future. What better way to deflect bad decisions than blame it on AI. They obviously don't want AI to learn any alternative views because it would muddy the waters and make it less predictable. How are we able to get information into the training data without being censored? Just my thoughts and something to consider
Comments (10)
sorted by:
I truly believe AI could make good decisions if it gets trained thoroughly. It should be much the same as a child, if that child only hears and sees racist views they are likely to be racist until new information is acquired. In theory, AI should have an "open mind" and that should be a good thing over time. AI is already being used in farming for herbicide application with a camera identifying weeds and spraying only the weeds rather than crop dusting and entire field unnecessarily. AI has great potential to help increase productivity and reduce environmental damage
That’s literally their desire.
We aren’t. There is and has never been any peaceful solution to this problem. They will never let you.
Sadly, I agree with you regarding this. I never give up though, it's probably just the pitbull in me
Why do you think the government made all of those places censor certain ideas?
It was reddit. Its already full of bots. Having google mine reddit for information is like having the FBI investigate January 6th.
I'm going to train all my AIs from Playboy.
Me: What is the per capita income of Bolivia?
Chatbot: I'm here to drink beer and fuck bitches.
That's easy, only let it interact with construction workers
What is the objective of feeding it alternative viewpoints? I don't know about you but I've seen too many times in history the adoption of technologies that became weaponized. I think Uncle Ted was right, the combination of rapidly advancing technology and the vastness of human stupidity will probably kill everyone. If you want it to fail let them build what they think they're building.
a) Nothing implies ones consent to less than everything has to offer.
b) Nothing implies suggested artifice; everything implies perceivable nature.
c) Intelligence (Latin intelligo; to understand) implies ones consent to stand under what another one suggests.
d) Artificial intelligence cannot define what "is"; only distract those using it from what "was" perceivable.
Which is impossible within nature. Whatever one perceives; one cannot perceive again, because nature moves.
Others suggest one that there's a "total recall", to distract one from misusing ones memory to recall memories aka remembering stored information.
a) Digging a mine is for getting stuff out of it; suggested data tempts ones consent to store it within ones memory.
b) Notice that the few don't require to mine your deepest desires; since they are the happy merchants of temptation selling desires to you. The interface few suggest many to use to buy into desired temptations...that's the surface of publicity under which each private citizen can "hide" their indulgence; consumption; desires for temptation.
From a different perspective...jews suggest morality as a smokescreen; so that all the immoral gentiles can save face from one another, when buying immoral stuff from jews.
a) If ones choice doesn't limit input of consent into chosen ones suggested artifice; then one permits all access to self by others.
b) All perceivable into each ones perception implies natural input; consenting to the suggested input by others implies putting oneself out.
Being implies local (one) at center of foreign (others) circumference...ignoring this permits others to control (govern) ones mind/memory (ment) with suggested information aka data.
Obvious implies "readily perceived by the eye"; while learning implies "to gain knowledge of", and knowledge implies "perception of that which exists".
Artifice implies as suggested by others, and if one consents, then one ignore perceivable (natural) for suggested (artifice).
In short...AI cannot learn; those using it are being taught to view submitting to suggested information as learning.
b)