• You are currently viewing our forum as a guest, which gives you limited access to view most discussions and access our other features. By joining our free community, you will have access to additional post topics, communicate privately with other members (PM), view blogs, respond to polls, upload content, and access many other special features. Registration is fast, simple and absolutely free, so please join our community today! Just click here to register. You should turn your Ad Blocker off for this site or certain features may not work properly. If you have any problems with the registration process or your account login, please contact us by clicking here.

How does a self-learning AI differ from a human learner?

Miaplacidus

New member
Joined
Jun 28, 2021
Messages
26
MBTI Type
TiN
Enneagram
593
Instinctual Variant
so/sp
Sounds like a silly question. Of course, there are numerous differences. an AI and a human are fundamentally different, one can say.
The question here, is not to compare AIs and humans in what seems the most obvious, that will be no fun.

Consider this, a child, in the process of growing up, keeps learning and learns in various ways - from self-experience, others, and outside sources (books, the Internet, etc.). The child may keep learning throughout their lifetime. Learning, after all, can be seen as a lifelong process, never-ending. Through learning, what is previously unknown to the individual, becomes known. If the child is deprived of all available (human) learning opportunities, what might happen? There were children who were raised by wild animals, and they acquire survival skills from the animals and behave like them. Without human intervention, those children might live their lives like wild animals, from which they have learned most things. If the child returns to human society and gets into constant touch with human civilization, can he/she learn the human ways? Probably. Just because they have lived like animals, does not mean they won't understand anything human. Learning can help them restore humanity. What about a child that has grown in complete isolation, without the company of people or animals? They are probably ignorant, and their cognition might be severely underdeveloped. As soon as they leave the isolation, they might start learning as well. Through learning, they can develop their humanity and be capable of reason.

The point is - it is learning, learning in the human ways, that constitutes an essential part of our humanity. Otherwise, one could only take the human form, lacking the essence of being a human.

Think about an advanced AI that can teach itself to learn almost anything, everything. It keeps learning and acquiring knowledge. It is also capable of synthesizing ideas to think for itself, generating something new. The AI has no humanity and cannot be said as fully human, yet, in this case, it is able to learn as a child does, arguably, perhaps even faster. If, the methods of learning are similar, and the outcomes of learning can match those of average human learners, then, speaking in terms of learning only, what would set the AI apart from a human? By constantly learning, will AI eventually obtain human cognition, one of the essential aspects of humanity?

Going back to the question - how does a self-learning AI differ from a human learner? One day, there might be few differences and AI can learn as well as us. Does it mean that humanity, at least part of it, can be acquired? It seems that having a normal human brain is neither a sufficient nor a necessary condition for humanity (or, 'human-ness").

In a world where AI can think as well as we do, what might happen to us? Considering today's reality that many people are comfortable with everything being provided for them, have little incentive to think for themselves, and might be either unwilling or unable to learn, will the future still be in our hands? Or, are the vast majority of the world population likely gonna be replaced?

A final note - We tend to learn through trial-and-error, and it can be slow and prone to mistakes. It can be seen as a primitive way to learn. Without revolutionizing the human learning process, making it faster, more accurate, and more effective, it is possible that we will fall behind AI in learning, and that AIs might one day, know more and know better than us. We should never let that happen.

Just some random thoughts.

Inspired by:
 
Last edited:

Methylene

Now with more salt.
Joined
Oct 10, 2016
Messages
2,202
Enneagram
639
Instinctual Variant
so/sp
I'm gonna look for the source, but a slight note: there have been cases of kids raised in the wild or isolated that were never able to fully recover. The Genie kid is an example, but if I recall correctly there have been more, even a twins study. This is unrelated to your main point, but it looks like if they spent long enough away from "humanity" their brains weren't flexible enough anymore to take in the new info, so they couldn't get fully inserted into society.

Answering the obvious side: machine learning is supposed to be more factual. You train a neural network after feeding it a set of data. The few algorithms I saw so far were all statistics based. So a basic AI lacks critical reasoning, the whys, the what ifs, and feelings, of course.
Anyway, it's interesting how we're trying and slowly progressing in developing something similar to a mind only with the aid of maths.

Very cool article.
 

Miaplacidus

New member
Joined
Jun 28, 2021
Messages
26
MBTI Type
TiN
Enneagram
593
Instinctual Variant
so/sp
I'm gonna look for the source, but a slight note: there have been cases of kids raised in the wild or isolated that were never able to fully recover. The Genie kid is an example, but if I recall correctly there have been more, even a twins study. This is unrelated to your main point, but it looks like if they spent long enough away from "humanity" their brains weren't flexible enough anymore to take in the new info, so they couldn't get fully inserted into society.
Right. Feral children are rare cases and those we already know, are discovered. There might be more. Among the limited number of known cases, some of them moderately successfully reentered human society and learned to regain their humanity. Others, however, had difficulty identifying with other people. They were more comfortable with the same species of animals they grew up with. It is suggested that if a child does not attain language skills by 5, then they will never learn to speak. The use of language is a unique trait of humanity.

What do these cases suggest? - 'Humanity' may be a fluid concept. Having a human body and a healthy human brain seem neither sufficient nor necessary conditions for humanness. Humanity is not innate and is determined by nurture, by the external environment one grows up in. Social isolation may not be a critical factor, as a child may still grow up as a human even if he or she has few/no friends. The most important factor determining humanity, IMO, is to be in human society, to be among other humans, with/without actual interaction.


One is not born human but grows up to be a human. Humanity is acquired, IMO, which means that it can be regained, lost, or weakened. If the essence of humanity is intelligence and reason, then an advanced AI, through learning, can gain that part of humanity.

Perhaps it can be said that humans are not, at the very fundamental level, different from other species as we also have basic desires, such as survival, food, reproduction, etc. What sets us apart, is the advanced hardware - the human brain. But still, different areas of the brain need to be activated and used regularly, otherwise, they might go dormant. Being in human society is the necessary activation process, without which there will be no humanity.
 

yeghor

Well-known member
Joined
Dec 21, 2013
Messages
4,276
A final note - We tend to learn through trial-and-error, and it can be slow and prone to mistakes. It can be seen as a primitive way to learn. Without revolutionizing the human learning process, making it faster, more accurate, and more effective, it is possible that we will fall behind AI in learning, and that AIs might one day, know more and know better than us. We should never let that happen.

Why not? If they surpass us in knowledge, maybe they'll surpass us in wisdom too?
 
Joined
Sep 26, 2022
Messages
62
MBTI Type
ISTJ
Enneagram
9w1
Reminds me of the myth Romulus - the founder of the city of Rome - and twin brother Remus who were nursed by a she-wolf (though this may be a euphemism for something else).

And there was some news a while back about a "racist and sexist" AI and it was blamed on the people who made the posts it read and not the AI learning system which had no conscience or filter or whatever. Also that one image predictor that spat out something like 90% of pictures of top halves of women predicted their legs would be bare because there's so much of that on the internet . . .

In a world where AI can think as well as we do, what might happen to us? Considering today's reality that many people are comfortable with everything being provided for them, have little incentive to think for themselves, and might be either unwilling or unable to learn, will the future still be in our hands? Or, are the vast majority of the world population likely gonna be replaced?

Right now with social media and stuff most people are comfortable with one narrative or the most shocking headlines of the day. If you believe in the existence of the soul then the machine is no problem but the real problem is transhumanism and repression of truth / disinformation campaigns to favor some other agenda through bot farms, huge data centers, surveillance, wire tapping, etc. This future is here NOW and every industrialized government engages in these practices to some extent - and every other one dreams of being able to use them.

Why not? If they surpass us in knowledge, maybe they'll surpass us in wisdom too?
The danger is that the AI starts manipulating its operators and they don't realize it. Tiny, subtle suggestions on massive amounts of data that require another set of programs to verify. All this can be falseified or tuned to further some agenda whose original purpose might have been forgotten.
 

The Cat

I'm from Outer Space...Dont Overthink it.
Staff member
Joined
Oct 15, 2016
Messages
19,387
One understands it is an artificial intelligence and adapts accordingly.
The Other believes itself to be the dominant species of an entire planet and acts out accordingly.
 

chubber

failed poetry slam career
Joined
Oct 18, 2013
Messages
4,406
MBTI Type
INTP
Enneagram
4w5
Instinctual Variant
sp/sx
The danger is that the AI starts manipulating its operators and they don't realize it. Tiny, subtle suggestions on massive amounts of data that require another set of programs to verify. All this can be falseified or tuned to further some agenda whose original purpose might have been forgotten.
I think this is what the Google guy who got fired (for saying it seems sentient) tried to say to the public, that these decisions, are made in a closed room of people deciding what ethics are.

Which also makes me wonder, if language is a living thing, and we have machines, that will be locked into a language accent, if you will. Will the children of the future form new accents? Or will this only happen when society goes to Mars and Google/AI of the accent that was chosen went with it, the children grow up with that particular AI, will their language structure and accent be similar to that of the AI?

If we read old English, can you make sense of it? The sentence seems different, but over time it evolved, and scholars also decided to change the spelling. What is going to happen in the future with Language? Will we stagnate and leave it to the AI?
 
Top