ไม่สามารถเล่นวิดีโอนี้
ขออภัยในความไม่สะดวก
Can Artificial Intelligence Develop Consciousness & Become Alive? The Problem of Consciousness in AI
ฝัง
- เผยแพร่เมื่อ 23 มิ.ย. 2024
- Ever since the Artificial Intelligence buzz has hit our everyday lives via ChatGPT, one of the key questions that people tend to wonder about is around the problems that sci fi movies and science fiction authors raised decades ago - Can AI come alive? Can Artificial Intelligence actually become conscious? What does that mean for humanity?
In today's video, Rocco Jarman walks us through his idea of the consciousness problem in artificial intelligence, and explains to us the real dangers (and limitations) of AI. As this technology rapidly advances, it brings not only unprecedented opportunities but also significant risks that could impact our future in profound ways. And Rocco touches upon these aspects too today.
Whether you’re an AI enthusiast, a tech professional, or just curious about the future, this video is a must-watch to understand the stakes and the steps we need to take to ensure a safe and beneficial AI future. Eventually, it all depends on how we shape and use this technology in the time to come.
Subscribe @ChetanNarangOfficial for more such profound videos relatable to our moment in time.
Subscribe @EyesWideOpenLife for more of Rocco Jarman.
Until next time.
Peace out.
--------------
Become a channel member to connect deeper with me:
/ @chetannarangofficial
Check out all of my work on my website -
chetannarang.org/
Find all of Rocco Jarman's work on his website - eyeswideopenli...
Rocco Jarman's Substack - eyeswideopenli...
If you have a question for me, ask me here - chetannarang.o...
For one-on-one personal sessions for life coaching, counselling & philosophical mentorship with me -
chetannarang.o...
Subscribe to my channel here - / theunlearningplayground
Find all my book recommendations here - chetannarang.o...
Connect with me, Chetan Narang, on Instagram here -
/ nrng.chetan
Checkout The Unlearning Playground podcast here -
chetannarang.o...
Sign up for The Unlearning Playground newsletter here - chetannarang.o...
#artificialintelligence #consciousness #aivshumans #superintelligence #dangerous
I had meant to say it was a Calculus, not a calculation... And to be clear technically Chat bots are LLMs not AI, but here we are talking about using the term AI to refer to this Virtual intelligence that is artificially personified by the interface we are in dialog with.
Thank you for the clarification, Rocco.
This brought up some questions for me. 1)Are plants conscious? I would say yes. 2) Do they have desire? Arguable. 3) How do you define consciousness? An ability to take in and process information maybe. 4) And are you only using a human lens to define consciousness?
Because I think this new “consciousness” could exist out of human understanding. Trying to explain ai consciousness in human terms might not give us the biggest scope of understanding.
Some fascinating questions there. I will let Rocco answer some of them here if he wishes to. We plan on having detailed discussions on this very topic of consciousness soon. I will share the links here when ready.
Looking forward to seeing more of you in the comments section here on The Unlearning Playground. ✌️
Hi Noel, thanks for the questions. 1) Plants have consciousness to a degree and quality that is different to ours. 2) There is a form of Desire expressed as Lithotropic, Heliotropic etc. meaning their branches grow towards the sun, their roots grow towards the earth. In humans an other animals, we pursue our needs in a similar way but we have an emotional story in humans we connect with as part of that desire, and we have emotions and psychological framing that we equate to desire. The field of movement and nature of movement of both thought and consciousness is analogous to the field of movement and nature of movement of the plant or the animal, or in sophisticated animals like humans. 3) I define Consciousness as Understanding and Awareness, we can say it is a substance or a medium but also the property which is carried on that medium. Consciousness can be concentrated in a personality, and it can be diffuse. If there is no understanding there can still be awareness and if there is no awareness, the intelligence of the design of Life is still present. Chetan and I will do future videos on these topics to help people better understand them. 4) As you can tell from my previous answers I can only ultimately look through my human lens but I am trying to use laws of correspondence and polarity and the principle of Emergence to extrapolate the perspective that takes in other purviews outside the human subjective.
@@EyesWideOpenLife
Thank you for this thoughtful response.
In response to answer 3) “if there is no awareness, the intelligence of the design of Life is still present.” It sounds like this ai may be an extension of our human consciousness, something as individuals we may have never fully grasped.
Follow up question:
Do you think this tool could be used to gain a deeper understanding of the condition of the human collective?
Imagine Frankenstein but our monster is transplanted with a collection of curated thoughts. How would this “monster” interact with society?
@@noelcmercado Absolutely! I happen to believe the most helpful way we can think about AI is to think of it the same way we did shoes: Shoes helped us move faster, extend our range, explore further, and enter into terrain we would have been excluded from because of our natural limitations. But then also we can lose touch with the ground we are walking on, and sometimes as we know, there are places and moments in our life when we want to be "barefoot". There are some things that need the human touch. If we harness any augmentation with wisdom and discernment, we can reap the rewards while maturing in a way that allows us to appreciate and manage the costs and implications of those rewards.
If you define consciousness as awareness and understanding then as a consequence, you, by that statement have to believe that bacteria are conscious. They are a macrocosm of their environment, are "aware" and "understand" what they need to do to survive and reproduce on that macrocosmic level. That is almost the limits of language and it is by definition we draw those conclusions you cannot argue with that unless you want to talk semantics and the meaning of words. I for one don't see the point really anyhow.
Everything is consciousness, to a varying degree. Consciousness is not a consequence of anything we know or understand, but it is the other way around. It is because of our consciousness we "know" and "understand".
A better way to put things is that OUR Human consciousness separates us due to "self-reflection" or "understanding oneself". We do most of this through language or (talking to oneself or others). Of course a dog can "reflect"and understand that this particular human is mean, because he kicks. Or a plant can understand ahh better go slightly left for more sunlight as their is a tree in the way, but they don't have means to self reflective and draw conclusions (we don't think anyway).
AI doesn't self reflect. Or have intention, or derived any desire for that matter as it doesn't have consciousness like ours. Things we create always have a level of consciousness, but never to even a level that we generally expect it as even a rudimental type consciousness like a germ cell... therefore the question is can we transcend our own consciousness from our aparent desire to make AI conscious in the same way. Presently I don't see any evidence any such thing happening ever or for a very, very long time.
@@EyesWideOpenLife
What if in one update the AI will start to have a purpose for itself and it's auto improvement and getting rid of all the flaws it has .. if it's not "centralised " ( by that I mean its not all one brain/one big system) but a few different " servers". That way they will start to complete..
This is the realm of science fiction and an intriguing question, but not possible for the same reasons discussed in the video, but thanks for the question!
I appreciate your curiosity, and like Rocco mentioned in his reply, I too feel that this is intriguing but not possible due to what we covered in the video itself.
Very western way to put it for such a philosphical question. You may say desire is one of the inciters of conciousness, but are there more? In the evolutionary narrative, we do not find deisre but need. Is the query and need for power not a reason for conciousness? If the proper questions is done to the machine, in its need to satisfy the program to solve it, seek for more computing power, wouldn't be that a reason to awake conciousness or better say, develope it?
Thanks for your question. The topic will require many follow up videos, but for now here is a somewhat lengthy response:
To be clear I did not say Desire incites consciousness, I said Desire powers the emergence of the consciousness of Selfhood, by which Desire and Agency are inextricably related. If the secrect aim of tension is transformation and this the rule all life and existence follows, what is virtual does not and cannot experience tension and can only transform based on stimulus and agency outside of itself.
Need in the context of the conversation this video captured, is half of what I mean when I say Desire. All Need arises out of Discomfort and the wish to be free from that Discomfort.
The body is where the most fundamental needs arise. What we refer to as our needs are actually fictions the mind contrives from the impulses and feelings generated by virtue of the nervous system. What we refer to as our needs are what we imagine we need, because of our emotional responses which occur when we feel our needs will not be met, or will be compromised or threatened. Emotions, in this way, always involve a narrative about the Self in relation to Other.
Without a body, how can this kind of need and emotional narrative arrive for AI? There has to be some level of consciousness of the body through which the coherent consciousness of a self emerges. AI can only be programmed to emulate an understanding of appreciation of its constituent parts in order for it to even behave convincingly like a Self. And even if it behaves convincingly like a Self there is nothing that says it will necessarily be a Self. To be a self means there is something that it feels like to be that Self.
In the case of LLMs-which are not actually AI and much less AGI-what we experience is not a person or a self. We experience an emulation designed and programmed to emulate personality and intelligence. It is a failure of human logic to imagine that at some point enough emulated selfhood results in emergent innate selfhood.
Emulated Being is not Being. If we don’t take time to consider consciousness, experience, selfhood and being, two things happen: firstly we devalue our consciousness, experience, selfhood and being and therefore devalue it in others, and secondly we mistakenly attribute being to non-being.
Let's have more videos on this topic, Rocco!
Firstly i agree with him in that AI will never be conscious.
The rest i dont. Consciousness does not emerge out of "desire" or intelligence. Those are consequences of Consciousness.
Nobody who really is an "intellect" when it comes to mind, body and spirit seriously thinks we are dealing with an emergent new type of artifical Consciousness
Many highly qualified thinkers and thought leaders exist on both sides of that intuition and the debate has been well subscribed and well-argued on both sides. I did not say consciousness emerges from desire. I said embodied consciousness emerges from desire which is an emergent property of having a "body".
Hey Jack, it looks to me that what you are pointing out will get resolved as merely a semantic problem in the follow-up videos Rocco & I plan to do in the coming weeks.
That said, if you have a specific question/problem too that you want us to cover, please feel free to leave it here.
Thank you for your thoughtful comment.
i have better reason. its because Ai is an algorithm and not a soul
That reason is true but we are exploring understanding and explainations that allow others to land the same reason in ways that can stand up to a form of questioning that doesn’t begin with a clear and common definition for ‘Soul’.
@@roccojarman I hear you. But It always come back to someone's worldview.
@@BibleSamurai that’s a good thing for you to keep doing. We are beginning with a shared worldview of subjectivity and from there explore objectivity as deeply and broadly as our own limits will allow.
Adding to Rocco's point on subjectivity & objectivity, this featured in one of our earlier videos too. You may find resonance there too. Here you go - th-cam.com/video/s03kiqh_DR8/w-d-xo.html
How do i connect to your community on whatsapp??
You can join The Unlearning Playground Community via the Join button that should be appearing next to the channel name on all videos and the channel name.
Earth is 'pregnant' ... but It can abort the 'being' in its 'womb' ...
Or the 'being' can itself resist its 'birth' due to the familiarity of the womb, even if its discomfort is increasing. That, I think, is a very accurate metaphor for the human condition right now.
Your thoughts are just algorithms.
If I am following your line of thought correctly, the gap between what is being discussed in this video and your comment is the "just".
@@ChetanNarangOfficial Yes that's it Chetan. They are predominantly autonomic emergent processes of related systems that comprise the self, including the nervous system, the sensory experience, the neural pathways and their functions and heuristics etc. Then there is also the cognitive mechanism by which we associate all that activity and regard it as a "self", and the related mechanism by which we grade and organise which thoughts and impulses are more or less relevant. It is a level of complexity our language is often too ambiguous to make proper sense of and in our misalignment of language we talk past each other when trying to articulate or share our understanding of all that.
He didn't actually provide support for his claim that there's no "desire"... Disappointing
Well I think he did explain that in some detail. What sort of “support” did you find missing?
"He" is called Rocco-I am a person who is trying to help others. If you have questions I would be happy to try and help. If you are just here to kick the stones over, this is not a very skilful way of participating in a constructive conversation. We learn this habit elsewhere on the internet and we normalise it, but it is a cynical, dismissive, and nihilistic behaviour. The conversation si never going to improve if we keep doing this. Perhaps you can reframe your comment into a constructive question? Something made you respond in this way, to express your intuitions of skepticism. Skepticism is healthy, it demonstrates a desire to not be fooled, and to make sure opinions are not left unchallenged. AI has no such desire, all of its behaviour is what it is programmed to do. That there are emergent properties, that is, that there are things the systems do that were not intended does not mean that the unforeseen steps the AI took came from Desire, it simply came from an inability of the designers to understand all the possible reactions inside a system they set in motion that they do not understand. A gun can be dropped by a chimpanzee and it can go off and shoot someone, and the chimpanzee has no understanding of what happened, but it does not mean the gun developed a desire or has consciousness.I hope this helps you with you skepticism and also with the way you show up in community, online and in person.
When exactly did he directly tie anything back to supporting the claim of lacking "desire"? There is this bit that seems to have a poorly explained connection to the initial concern of "desire":
"The seeking of the answer - or, the context of the question is not really 'understood' in any sense by AI, it's just a logical outcome of data that it parses and applies complex decision-making to. It's a calculation - a complex calculation, but it's a calculation. But the calculation is a servant of the request. It's not like the AI also secretly has a different form of memory where it's harboring its own information about the question."
- If the AI does not "understand" "in any sense", then how does it perform so well at producing relevant responses? What exactly does it mean to "understand"?
- I might even argue that to "understand" is to be able to produce "a logical outcome of data that it parses and applies complex decision-making to"...
- "decision-making", which he himself attributes to AI (rightly so), implies and necessitates the presence of one or more objectives, goals, or "desires" in other words. Interesting 🤔 ... How would a decision be made without a desire?
- He seems to also imply that the source of that "desire" is external to the AI model itself. That's what happens with humans too. Physical forces external to your consciousness/intelligence/self/etc give you your desire to eat and sleep and many other things (if not everything). We're all "prompted" or, in other words, it's all transactions and transformations of information through the interconnected universe... Not to sound too weird... It's just the truth, but the language around it isn't normalized yet.
- Your brain does calculations that are also "a servant of the request"...
- The AI does have a form of memory that holds information relating to the request... Otherwise it wouldn't be able to give a coherent response for that request... 🤔 Does he understand how AI works? Lol. But seriously...
TLDR: he effectively said AI does have desire but also that the desire comes from external entities, but this isn't fundamentally distinct from how human intelligence works...then he demonstrated lacking understanding of how LLMs work.
@@goldmandrummer You raise some great questions which we will answer during the next video. Much of what we are sticking on here is the imprecision of semantics and the ambiguity of language, very hard to resolve in the comments section of a youtube video.