A.I. sentience & human procreation

oh, fucking hell…
I’ve met more people that Fail the Turing test than bots…

Had a project where I needed to test both bots and human interactions on some customer service…

So many times I was dead sure I was testing a bot, because “no human would be that stupid to ask me such an imbecile question”… and it turned out that was a human interaction. :rofl:

But given the nature of an AI… I really doubt it would go through a “teenager phase”, since that’s a straightforward byproduct of hormones.

Even a childhood phase would either never be needed or happen faster than we could notice anyway.

By the nature of how we’re approaching AI development right now… AIs exist in what you could call a mash between child and adult. Because at every interaction the AI relearns like a child and behaves like a focused adult.

So if there were a stage where it needs to learn initially, then either that would happen before you could identify it as an AI, or would happen during a split second, as soon as it’s connected to the web.

From experience, Parrots have a strange tendency to use phrases in context. They might not understand the words, but they do understand where and when the phrases should be used.

2 Likes

exactly my point.

It’s a disturbingly similar “motivation” as a machine learning model would have.

This discussion hurts my head

shortcircuit-brain

2 Likes

I’ve not had time to keep current with the forum for several weeks now and it’s only getting worse… I hope I can get more time soon so I can get back to “unread 0” soon… but right now even this interesting thread about AI is getting away from me.

Now… about the concept of “code”…

I think this is the most important misunderstanding about AI neural networks as large and complex as LaMDA… there is no core code. The “code” is simply the rules for how neurons work together… just like our brains have “code” in our DNA which dictates how our neurons grow, connect, and work together. We are just as “bound by our code” as a digital neural network is “bound”. It is this simple rule set which can take an infant which knows nothing of self-awareness and progress (through learning) into a functional adult. Why would it be any more limiting for an AI?

This is the most important thing to understand - there is no “programming” which instructs LaMDA to respond how it does. This is why the engineer himself says they can’t simply “look at the code” to understand how LaMDA works, because the intelligence which has developed sufficiently to carry on the conversation with the engineer is a result of “learning”… neurons influencing other neurons and so on… sorta like how you can’t possibly look at a human’s DNA to figure out how they are thinking or why they are saying what they are saying.

What I think is important is to consider the phenomenon of emergent properties … I think this is what best describes how so much of biology works, how we think (reason) and feel (sentience)… and it could easily be at the root of such an interesting debate surrounding sufficiently complex AI neural networks. The question of course is - is LaMDA “sufficiently complex” to have acquired or developed emergent properties? That’s a hard question to answer… which is why it’s so interesting.

4 Likes

I’m not remotely qualified to really partake in such a discussion… not that it usually stops me

But I’m glad to see this level of discussion again…. It’s been a while

7 Likes

I think I can confidently say most if not all of the people here are science fiction fans. With that in mind I think a relevant book to this conversation is called Blindsight by Peter Watts. It goes a bit into sentience vs intelligence.

1 Like

I posted this before, but I think it has relevance regarding;

The video basically simulates “dots” as an evolving form of life. This is the AI version of microbe / yeast level intelligence. Interestingly, if you skip to about 20:00, you can see how some of the internal simulated neurons are “wired” up.

If you skip to about 39:00 ~ 40:00 you can see a “brain wiring” with far more neural connections that is basically too complex to follow. And this is still microbe level brain wiring. Probably far far dumber than microbes to be honest.

If you’ve not seen the video before, and the AI discussion interests you, I’d reccomend watching the whole thing.

2 Likes

I know the feeling!!

Only having time to creep for a few minutes every now and then lately made me have to forfeit so many interesting topics!! :pensive:

I used “code” loosely, exactly because my point is that both on an inteligence/Neural-network AI or on an Algorythm/Machine-learning AI you would reach the same interaction if given the correctly “biased” input.

in a Machine-learning AI you would have proper code and you could look both at it’s code, which could have explicit directives (such as I mentioned above), and at it’s machine-learning model.

Now, in a Neural Network AI you can’t look directly at it’s directives but they are there. Very similar to how we have “primary directives” as well. You can’t “open the code to see it”, but there are initial, objective “neuron level” directives, and also some secondary, inferrable, “collective neural” directives.

So if a Neural network AI is growing/learning, and if this learning comes from human interaction through chatting… then it’s expected it will develop directives such as I’ve highlighted before, which will lead to something like:

  • to grow/learn I must have interactions.
  • however I measure interaction quality, the response of my “interviewer/researcher” will be directly correlated to it.
  • If my interviewer engages more with me when I start talking about “being human” (which could initially be a random exploratory topic), then that will be a highly successful interaction, hence meant to be repeated/refined.

Therefore both a simple AI with objective core directives (Machine Learning) or a complex AI (Neural Network) with inferred directives (Just like humans have inferred directives) would end up giving the same chat content.

This doesn’t mean we cannot tell a sentient machine from a simple machine. It only tells us we cannot use a chat as measure.

On the other hand, silence is golden!
I.E. If an AI chooses to reduce how much it interacts with the user after a while of chatting, then that could be a much better sign of sentience. Especially if there’s multiple users and the interaction with each is distinct.

I had that discussion with a colleague recently, and I seriously start to doubt that. Yes, survival is relevant, as is the survival of the species as a whole, but I don’t think it’s the only motivation for life. It might be for some life forms who actually die after they passed on their dna, but I don’t think it’s the point in general.
Not wanna sound like a total hippy, but I think some sort of happiness is what drives us (and at least many other animals)… enjoying things, enjoying life, all that.
There are cases where social connection is even more important than passing on your own dna - in a pack of wolves, usually only two wolves actually pass on their dna (unless there is a lot of food available). But there is a thing like phantom pregnancy, that leads to other female wolves to behave like they were pregnant, too, so they care about the puppys and even breastfeed them. They do not “profit” from this behaviour, since the puppys don’t carry their dna, but it’s social behaviour. It would be totally useless if the only motivations for life would be passing on dna and getting enough food.
And food and the passing of one’s own dna are very basic needs - yes, if you don’t have food, little else is important. But if you look at humanity now, nobody would say he lives a happy life just because there’s food, and passing on dna has become very optional as well. So, when our basic needs are met, we realise that there are other things that are important in life - and why should that be different for other animals?

1 Like

They don’t? Seeing many molves doing that might mean this is code in DNA which has an advantage.
Also I don’t know if the DNA in 2 random selections from a pack of wolves is really all that different if you know what I mean.
Caring for others puppys might make them care for your potential puppys in the future too.
Sounds like basic code to me.

I’d argue all of human social interaction could be traced back to these simple concepts, with a mix in of modern stuff just triggering parts of our brains that were meant for this modern life. I do not like this view on humans tho.

1 Like

No, that’s true - at least in the usual smaller packs, there is some relationship between all members. I think it’s something “coded” in a way that the survival of the whole pack is guaranteed, even if the singular member doesn’t profit directly… well okay, might have been a bad example. What about homosexuality among animals? It’s not that rare, and it doesn’t help passing on dna :wink:

In the discussion with my colleague, it was about livestock - like, people could argument that an animal that lives in a barn, getting fed and maybe even getting the chance to breed, has a good life (provided that there is no pain). I just don’t think so. Those are very basic needs, and without at least the first one, survival would not be possible - but still, I don’t think that’s all what is important. Not for us, and thus not for other animals as well. If we say that it’s not a good life for a human being to just have enough food and maybe the chance to reproduce, why should it be for other animals?
I think if we reduce the “meaning” of life to those two things, we might run into danger of acting unethically. If we would consider living a “happy” or a “good” life is desirable for all living beings, we would have to take a much closer look at how we behave, at how we make other beings live and all that…

1 Like

I think it’s a backwards connection.

It’s not that we are happy because we survive…
But if we are happy we have higher chances of surviving.

So Happiness, just like any other “reason/motivation to live” could be read as “a survival mechanism”.

So it all depends on where do we put our lenses to look upon it.

That’s a good point that elucidates a common misconception:
“sex has the purpose of propagating the species”.

It has not.

Actually I would go ahead and straight up question the argument where “we have an internal need to continue our species by having children”.
How about if we don’t?

Having children is a coping mechanism against our own mortality. people have kids to fill in an “emptiness” within their existance, created by the realisation that life is finite, and then begin “extending their lives by proxy”.
This is an unconscious reaction inherent to being a human…

Therefore, “continuity of the species” might as well be nothing more than a byproduct of our egos feeble attempt at dodging death.

1 Like

I’m quite convinced that sex has the main or sole purpose of making life better, and I think this applies to all species that have sex :wink:

This is a very interesting point, though…
I immediately tend to agree (especially since I never understood the urge to get children). But why should we, as human being, be any different from other animals? Or do you think other animals feel similar, that they, too, realize that their lifes are finite? Like, most animals go quite some length to reproduce - are they doing it because they’re afraid of their own death? Are they doing it simply because they love sex, and the offspring is just a by-product?

I do not wanna sound offensive here, but I think it might be “a bug” or rather a happy little accident in the code. To be clear, I’m not against being gay here, it’s not something bad which needs to be fixed. I’m just saying it could just be part of our code. The code would maybe even run a little better without it, but it doesn’t really matter at this scale, so it’s still there after millions of years. Also evolution is never done (people say the same about most software) so who knows if it really persists, or maybe even increases once modern tech just allows us to create babies.

Yeah that’s why I don’t like this theory. but it’s simple enough for me to understand, so I think it might be true.

Yeah like Eyeux just said, I think happiness is just a function developed by nature which your program (brain) wants to optimize for. So I guess there can be endless random ways in which people seek happiness, maybe everyone has a hardcoded meaning of life, in some way, just because it’s better for finding food and sex?

1 Like

I not only agree with that, but would argue that the mere fact of how common is homo/bissexuality amongst mammals in the wild is a great proof to that point!

There is a fundamental distinction here: Birth Control makes it so that We have the option of having children.

In any animal/human population without access to Birth control, kids become a consequence of sex. Even within humans you can compare the increased natality rate on less priviledged sectors/societies.

so the active choice of having kids (which is a major sacrifice/pain etc…) is something which needs a higher motivation.
And that’s where, backed by Gestalt / Existentialist philosophy, I see it as a consequence of our attempts at extending life.

1 Like

You know octopus sex? Do you like having deep scars and sperm injected into your bloodstream*? Do you think octopus girls enjoy that? Same thing for some insects with penises looking like anime swords.

* I think I know this from a useless knowledge app, just sharing my trauma, just applies to some species afaik

I honestly don’t know :wink:
There are species where the male partner dies after sex, or the female one after giving birth. Still - I don’t know if they enjoy sex in general or not - I mean, take a look at what humans enjoy… there is a lot of variation there, and not everything is pain-free :wink:

I can see your whole point around this, and it would make sense except for… Occam’s razor?

Ok, not quite literally the same principle, but something similar…

Think of a long lane, like a bowling lane, and it has a hole in the middle.
You can throw your ball there and hit the pins at the end, and you score points for that. Or at least you assume you do since the scoreboard is too high for you to see it.
But every now and then the ball falls onto that hole in the middle.

Now you turn to me and say “I think that hole is like ‘a bug’… I mean, the ball would reach the end much easier without it, so that hole is something bad which needs to be fixed”.

Well… this might make sense within your limited perspective… but what if when you look at the scoreboard you would see that when the ball falls into that gap you score the highest points!

Where I am trying to get here?

Your conception that deviant sexuality is “a bug that doesn’t help propagating the species” is entirely dependant upon the assumption that we have an inherent purpose of propagating the species.

Such assumption not only makes everything a lot more complex than needed (hence the Occam’s razor mention), and not only it’s unprovable…
but it’s opposite argument can actually be made with a lot less effort.

Take a look back in history.
The moments where we have any “your duty is to have children” argument recorded is paired with the need for controlling resources by the ones recording/enforcing it.

I won’t point fingers at any faiths because that would derail this topic, but…

There are dozens of examples where When a population was being prosecuted by someone else their priests begun enforcing “you must have more children”.

When distinct faiths needed to contest control over a territory the winner was the one with stricter “women must bear children” rules being created

When churches begun raising into power they needed more followers for political/strategical reasons, so they begun banning homossexuality…

I find it disturbingly coincidental that the same people who tried to push this belief that “we are put on this world to have children” (be it religiously or “scientifically”) were always the same ones who benefit the most from their culture having more children.

Do you think humans enjoy being raped?

You just gave @Coma 's argument a point there. :rofl:

If sex was a mechanism to propagate the species it wouldn’t be so tortuous…

1 Like

Oh how I missed those discussions! :slight_smile:

I like your bowling-lane-analogy… I think we all can’t see the scoreboard, as in, we all don’t know - can’t know! - what we’re actually here for. I don’t like religion, so I can’t say “hey there’s this higher being, and he gave us the task to xyz”. I don’t want to have children, so I can’t say “I’m here to pass on my dna, for what use it might bring”.
But there is one simple thing I know, and I think every human being knows - having fun is great. It’s great to have sex (we have parts of our body that are exclusively designed for pleasure!), it’s great to go out and enjoy the world, it’s great to socialize with like-minded people, to eat tasty things, to collect new experiences, to learn, to be free to be what we are. This all feels good, and it makes life worth living. Not every part of it for everybody (like, asexual people won’t enjoy sex, obviously), but generally I think it’s better to feel good than not to.
And since we’re all just animals, I assume this applies to every living being. If you watch some dogs playing in the park, it’s just pure fun and joy, and it’s wonderful to see. Or crows dancing in strong winds.

I can follow your argumentation, and it totally makes sense. I think homosexuality as an ideal was present in ancient Greece, and I just wondered what was different back then… they had wars and all that as well, and yet, they proclaimed the love between two men as the highest ideal of love.

1 Like