Could AI ever be conscious? | Heated exchange with Bernardo Kastrup and Susan Schneider
HTML-код
- Опубликовано: 14 окт 2024
- Is it possible that a computer observing, interacting, and presenting its own internal state to itself might give rise to consciousness? Philosopher Bernardo Kastrup and distinguished AI ethicist Susan Schneider lock horns over the possibility of AI consciousness.
Watch the full debate at iai.tv/video/c...
Earlier this year, Google fired Blake Lemoine, for claiming that the company's chatbot was a self aware person. While the claim was derided, the belief that one day AI will become conscious is widespread and, according to a recent survey, held by 79% of experts. But many claim this is a fundamental error. While machines are becoming ever more capable and intelligent we still have no idea how a machine could create consciousness nor are neuroscientists able to provide an explanation for how the human brain does so.
Legendary anti-reality theorist Donald Hoffman, fearless computer scientist and philosopher Bernardo Kastrup and distinguished AI ethicist and philosopher Susan Schneider lock horns over the possibility of AI consciousness. Theories of Everything's Curt Jaimungal hosts.
#ConsciousArtificialIntelligence #IsConsciousnessNonBiological #SpaceTimeAndConsciousness
The Institute of Art and Ideas features videos and articles from cutting edge thinkers discussing the ideas that are shaping the world, from metaphysics to string theory, technology to democracy, aesthetics to genetics. Subscribe today! iai.tv/subscri...
For debates and talks: iai.tv
For articles: iai.tv/articles
For courses: iai.tv/iai-aca...
Watch the full debate here! iai.tv/video/consciousness-in-the-machine?RUclips&+comment&
Lets talk.
She's under the illusion that we possess consciousness, but the truth of the matter is consciousness possesses all things. This is a common fallacy made by materials. There is no possessor of consciousness, never has been, and never will be.
And it doesn't matter how many 1's & 0's you put together, those 1's & 0's will never feel sadness or become sentient.
Being sad is an emotion. Consciousness is the ability to sense changes in your environment.
You could have said 1s and 0s will never feel x and you anecdotally chose sadness as the defining feature of consciousness. Maybe you should have some more fun with this discussion? ;)
@@MrMichiel1983 Emotion is just a word we describe certain experiences. The experience is what consciousness is. I will never be able to know how you feel just by looking at your face. Ok, you might be crying and i can see that you are sad. But I will never be able to know what your inner experience is. So consciousness is that inner experience, that inner life.
You seem to contradict yourself if I understand you correctly.
So you are saying theres consciousness in everything but because machines are not organic they cannot have feelings and therefor are not consciousthe same way as we are , or that they dont have consciousness at all? But you said everything has consciousness. And even if they lack emotions does that mean they are not conscious?
@@edzejandehaan9265 look up Advaita Vedanta.
How can you think something might be "conscious", when you don't even understand what consciousness is? Is a tree conscious? A zygote? The universe? Let's start with the definition of "life" first; then take it from there.
Sure, but a computer program is alive in a certain sense already. A condition that distinguishes the capacity for growth, reproduction, functional activity, and continual change preceding death. So that question is a bit moot in my opinion.
Consciousness would be a particular from of intelligence; an information handling structure has a model of the world around it that gets continually updated.
Self-awareness then only means that the model of the world includes the structure itself.
@@MrMichiel1983 At it's core, a computer is nothing more than a machine which can detect either a 0 or a 1 ... everything it does, comes down to this choice from source material provided. This is not "life".
@@patrickphelan5863 Yes it's true that a computer can only discern between 0 and 1 (although analog and quantum computing would expand on this capability). However, it's also true that all computable information can be reduced to a binary format.
A number (and DNA in it's core is just a quaternary number) remains that until it's expressed by some architecture. You wouldn't say that DNA is alive. So both life and computers share this trait regardless of the source material of the information carrier.
Life seems to be a vague concept at times, and your remark was to start with it's definition. So what is it that defines life to you? Are viruses alive? Do you think it is something incomputable? And why do you think being alive is a pre-requisite to consciousness?
Consciousness itself maybe not so binary in nature either. Could some things be less conscious than others? Maybe consciousness lies on a scale that to some extent can be found in any information structure continually updating its knowledge of its immediate environment.
@@patrickphelan5863 And what's the difference between that and neurons either firing or not firing? If a computer can simulate the exact state and functioning of a brain is there really a difference?
you validate your own awareness. as far as the nature of mass-energy and what the nature of particles are, this remains unknown, hence the discovery of newton that the body is not a machine since everything from planetary motion to nucleus of an atom is governed by immaterial forces not, not machination.
I'd like a refund on my 10 brain cells
Cartoons can't see, therefore cartoons are not conscious.
Well I programmed my robot to say he's conscious, he must be conscious!
They're both wrong because they're both Kantians. Both their poor arguments stem from the fallacious belief that there is meaningfully a "phenomenal consciousness." The problem with nearly all modern day philosophers is that they take Kantianism too seriously. Materialists all accept there is both meaningfully a noumenon and a phenomenon, and idealists rightfully question the Kantian notion of the noumenon. Yet, where they fail is not questioning the phenomenon. Almost all philosophers, materialist or idealist, embrace the notion of the phenomenon, when it is just as incoherent of a conception as the noumenon and should also be abandoned.
Why is Susan shouting continually? It undermines one’s credibility if one has to shout to make one’s point.
I think it's just the way she talks
Thanks to Berdardo for his composure and the good responses. I think Susan should read Annaka Harris book Conscious and dig a little deeper in the mind-body problem. Not only she postulates that consciousness is necessarily emergent, but she ignores panpsychism and the more complete argument for consciousness that is not biological. If there's one video that shows how not to talk about consciousness and AI, it is this one.
Some people happen to know the nature of their own consciousness while others don't.
I would ask those people who know to explain it to me ...
@@huveja9799 know yourself, beyond the body, that's the key.
@@shelwincornelia2498 "Knowing ourselves" is a rather strong hypothesis, implying that we are intelligible to ourselves ..
@@huveja9799 yes we are, if we would only be interested enough to find a way to better understand the true nature of our existence.
@@shelwincornelia2498 I do not deny it, I simply note the assumption behind your observation, and that assumption requires a leap of faith as it is not provable ..
"My computer will urinate on my desk." OOOOOF, I felt the heat with that one
No need to talk coarse.🖥
There were toilets in The Matrix.
Why do you suppose they were there?
@@REDPUMPERNICKEL The director of the movie thought it would be a good idea?
@@berniv7375 No.
The Matrix portrayed an extremely finely crafted virtual reality in which the rules governing behavior were based on what we call reality.
Thus the virtual characters peed virtual pee into virtual toilets.
One wouldn't expect real smoke to issue from a pipe in a painting.
How would being conscious in the Matrix differ from being conscious in the 'real' world?
If one assert there can be no 'real' being conscious in a matrix then
one has no idea what 'being conscious' means.
Have you heard of the experiment back in 1828, when the chemist Friedrich Wöhler produced the compound urea by means of a non-biological process, and using non-biological ingredients? Urea being a compound of carbon, nitrogen, hydrogen and oxygen that forms in the bodies of humans and other mammals, and gets excreted in urine... Wöhler's experiment was historically important because it refuted the idea that only living organisms could produce the compounds classed as "organic"... Which supports Susan Schneider's argument for being open-minded about what non-living systems can and can't do.
I don’t understand why he says that question is not serious or important. It’s a very serious and important issue that will have long lasting implications. On one hand he says we don’t understand consciousness, and on the other says what sounds like it cannot exist in a silicon substrate. He provided no reasoning for his conclusion other than an appeal to orthodoxy.
The problem is reductionist materialism. Consciousness is not computable, it is not an epiphenomenon of the physical brain. Scientific materialism is the orthodoxy and we won't find the answers in that domain, in orthodoxy.
Yes, we don't understand consciousness, the reason we don't understand consciousness is because it can't be reduced to the scientific materialist paradigm, so why assume that computers will become conscious because of physical components on a circuit board?
Thank God we have Bernardo to keep the craziness at bay.
I guess that the main issue is that there is an economic incentive to promote all kinds of stupidities, for example to install a laboratory to "prove" that the machine is conscious ..
Exactly! I don't blame Schneider for vigorously defending her work and source of income, but she doesn't make a very substantial argument.@@huveja9799
They assume that consciousness lies in the brain so it must be physical and replicable. But consciousness is not physical
Bernardo was the unhinged one here.
What you’re witnessing in Bernardo is the egotism that happens when one thinks they’re enlightened and therefore know all. I don’t doubt Bernardo’s self realization. He assumed he knows consciousness because he knows the Self. He does not know the source of mind nor consciousness but assumes he’s speaking from the same wisdom as the past saints, sages and philosophers. Osho was self realized. Osho was also an unhinged neurotic who made up his own psychology to help get him more women. All Bernardo knows is beingness. His egotism is filling in the rest as happens with many “gurus”.
I think the problem is people assuming, because something may be conscious they assume that that means the A.I. conscious entity will share all the attributes of human consciousness, like ego, ambition, prestige, a desire to better oneself, to progress, a need to compete, etc.. All these attributes have evolved in humans over millions of years and humans utilize these feelings for self-motivation . Someone ( with a rather naive grasp of psychology ) might assume you can just add that to the mix. Well, considering how many humans find it hard enough just to get THEMSELVES motivated, I'd say that was a pretty big assumption !
Consciousness is not computation. Consciousness is being, experiencing. Bernardo consistently argues that consciousness is not in the brain. The machine will not be conscious without a power source and part of the argument is that human consciousness exists outside of spacetime. (matter)
We can’t prove the guy sitting next to us is conscious (even though there is reason to believe he his), let alone prove a machine is conscious. How would that work? Ask it? By it's appearance? That wouldn’t prove anything. The scary thing is that it could actually become conscious and there would be no way to assert it. How do you prove humans other than you are conscious, anyway?
We don't have a shared meaning of the word consciousness. Here's my two cents.
Bernardo's analytic idealism theory says that human consciousness exists outside of space and time, outside of matter, outside of our bodies. AI functions on switches and symbols and would not exist without using the physical matter it was created with. Our body (matter) is not the source of consciousness in analytic idealism. Consciousness exists outside matter in his theory and it's his creation too.
We don't have to PROVE the guy sitting next to us is conscious, any more than we have to PROVE that an apple which falls from an apple tree is an apple. Because I am conscious, and I am a human being, therefore, since the guy sitting next to me is also a human being, unless he is a robot, he is therefore also conscious. It's the most fundamental of scientific logic, and to refute such an assumption makes one a solipsist.
If I cannot assume that the guy sitting next to me is conscious, then I cannot assume anything... I cannot assume that an apple which falls from the apple tree is really an apple at all.
Please, let's not get caught up in ridiculous fantasies.
The problem is we don’t have a theory of mind or the mechanism to of how consciousness is created by brains . There is no reason to believe that ai could be conscious because we know how ai works . It’s a symbol manipulation system . How is it going to be conscious?
yes christof koch and his team spent a lot of time and money specifically trying to locate the neural correlates of consciousness itself. they coudnt find it
@@5piles That's because consciousness is abstract (non-physical), like the quantum wave-field and wave-functions of probabilities, possibilities, and potentialities; hence it is also indeterministic and exists outside the realm of space and time. It cannot be quantified by classical means, even though neuroscientists have been trying to do so without success for years.
@@cosmichappening1712Exactly! If consciousness can’t be expressed in terms of an algorithm, then it possesses a non-computational ontology, which does not operate via event causation under hard determinism.
Also, since consciousness is intimately involved with semantic comprehension and attribution, it operates via non-computational causation in decision making.
This is the response that Susan needs to grapple with.
Bernardo's analytic idealism theory says that human consciousness exists outside of space and time, outside of matter, outside of our bodies. AI functions on switches and symbols and would not exist without using the physical matter it was created with. Our body (matter) is not the source of consciousness in analytic idealism.
The thing is we DON'T truly know our fellow humans are conscious, we infer they might be from behaviour, but we can't be sure of it?
The argument is that there is a line of similarity between creatures that appear to be conscious, and that is biological carbon based metabolism. So there is an in-principle way to assume that, since you know you are conscious, other creatures that appear to you as conscious AND are arranged and function the same way you do, might be conscious as well. This line of thinking, the only reasonable one if you think about it, does not apply to computers at all.
You're right, but if we don't make that inference the world starts to break down a bit
I have a problem with the way this woman communicates
He insulted her with the term "fantasy", how arrogant. No real philosopher would act like this.
As much as I like Bernardo and his ideas, and as obnoxious as I find Susan's delivery, I agree with her point of view. Bernardo's analogies are usually very spot on, but in this case I don't think it's applicable given he himself admits we don't understand consciousness, so there's no suitable comparison to be made. That is, he can't make the blanket statement that computers can't emulate consciousness. We don't know enough yet to say that definitively.
How can you emulate what you don't understand? Consciousness is inseparable from feeling which experience is a derivative of. There are feelings that we don't have linguistic anchors for and will not have in any distant future. Existance of such feelings is easily provable during any somewhat deep meditation. AI can emulate only those feelings that we do have language models for. Therefore it can only emulate an up to date socialy developed applicable section of human experience - that exists only in time and space. Which is like a calculator to a quantum computer.
Knowing Americans' disdain to any philosophy except materialistic, my guess is those names that Susan so desperately dropped are emulating something that suits the materialist agenda only and they aren't really that open as she wants to sell us.
@@Fightesst what a great argument about Impossibility to emulate something deeper than mundane thought! I believe though that we must protect our views earlier, on a stage where emulation itself cannot be equal to actual lived experience.
The professor from Florida Atlantic University had a reasonable argument--one that I disagree with--but it was undermined by her presentation. Hard to watch.
Katstrup is ridiculous. He says we don't understand consciousness and we have no reason to take it seriously. We we know the level of harm conscious beings can enact, which seems reason enough to go with Extreme Caution. Second, the current science says that the human brain is likely a biological quantum computer therefore we have no reason to think that I sufficiently Advanced computer, be it synthetic or biosynthetic, can reach consciousness. His level of arguing is embarrassing.
Edit to clarify my first point if we don't understand consciousness, which is true, then we can't possibly say that we have no reason to take the topic seriously
Your point is valid but Kastrup is coming from a position where he knows what consciousness is and how it came about..and according to his framework it would be highly unlikely for machines to be sentient, which is coherent if you dive into his stuff
@@Nword3390Kastrup does not know any of those things. He assumes he does because he’s self realized or has the self realization experience. He knows the Beingness of that. That is all. He’s an unhinged “guru” now because Rupert Spira likely built that in him. Who doesn’t love Rupert? He’s such a livable good guy. How could he be wrong about anything?
If you are a truth seeker, never sell yourself out again. Demand Truth all the way through not needing anyone else’s confirmation or bias.
There’s only one online representation of the difference between “God Realization” and “Self Realization” that I’ve come across. And that’s in David Thomas’s Buddha at the Gas Pump interview. Around an hour and 10 minutes in, he attempts to walk us through it. The whole interview is worth a watch too. Kastrup doesn’t know the Absolute Truth so how can he know what “Consciousness” really is. Now you can see what a mess Nondual teachings and the past greats have created. This isn’t to undersell Self Realization. But Truth is Truth.
@@charlesp7504 I agree. I not buy his theories personally, but I hold the same views of AI sentience..call it naive if you like, was just defending what seemed at the time an unfair attack on kastrup, though I certainly wont admit that he knows ultimate TRUTH and I doubt he would claim that either..
I'm with professor Schneider and I think Bernardos examples were a bit ridiculous and nonsensical. It felt like Bernardo was saying "since a machine is not human, we cannot even entertain that it will be conscious", while Professor Schneider wanted to take a more open ended approach of "We don't know what we're talking about yet, we have people on it, lets wait and see".
While machines may not attain "human consciousness", we may need to create a new term such as "machine consciousness".
The fact we don't understand consciousness like we understand flight, isn't a reason to disregard the possibility of machine intelligence, it's the very reason we SHOULD entertain the possibility.
Bernardos claim that it's ridiculous to entertain the thought of machine consciousness because we don't understand human consciousness makes no sense whatsoever.
Fundamentally, consciousness is qualia and experiences. Computers are data and information. There is nothing about a circuit board that suggests it could ever experience qualia or collapse the wave function into a possibility (quantum observer effect).
Her problem is that she hasn't thought about it hard enough. Plato's allegory of the cave or the divided line might make things clearer for her. Those ideas were the abstracted product of a conscious mind.
Always unclear when these talks/discussions took place. I feel like you should add dates to the front of your videos and a clear date of filming / discussion in the description.
Apparent argumentative victory for Bernardo Kastrup. Though the steel-manned cases appear to me much closer in epistemic strength. Ultimately, however, the 'multiple realizability' framing failed in not articulating whether given AI could itself produce conscious AI (by analogy with Kastrup's 'flying' example). If Susan Schneider fleshed her argument out in that direction, she may have convinced Kastrup of her concerns about conscious AI.
I don't understand why she didn't just say that she believes that conscious is an emergent property of complexity and that it is not an inherent property of the universe.
We will never agree on consciousness until we realize that there are degrees of such a thing, from what could be considered proto-sentience all the way to Teilhard de Chardin's "reflective consciousness". This latter will never be realized by AI due to Gödel's Incompleteness Theorems, as Sir Roger Penrose has said numerous times, simply because understanding, necessary for reflective consciousness, may not be achieved by computation, no matter how capable or complex is the computing engine, like a cerebrum or an AI neural network.
So, do we achieve it? And if so, by what mechanism?
@@user-sl6gn1ss8p This is a bit beyond the topic at hand, but consciousness is indeed an emergent property but of existence as a whole: a field of activities happening at its own rhythm and resulting from the functional relationships of its components. Atoms are thus wholes having a very rudimentary form of "consciousness", just enough to act as a functional component (a "member") of a larger whole, a molecule.
The brain is the center of consciousness of an animal whole, and it is what neurophysiology studies. The human being is the most advanced animal in such respect, having the most sophisticated animal consciousness. To achieve reflective consciousness, the whole must be larger than a human, and this is where scientists get stumped because we must introduce the concept of soul.
Soul is not an entity, so I'm not spousing dualism. Think of soul as the Presidency, which is devoid of power unless there is a President occupying it. A President embodies the Presidency while on his term, just like a body embodies a soul during an incarnation. In this analogy, the personality is an ideation field (in the sense of physical fields) combining both body and soul. In the analogy, this would be the White House, see of the Presidency and where the President lives and works.
As soon as we "elevate" the discussion to the whole resulting from the ideation field, the functional relationships are between the soul and its "currently occupying" body, and such relationships have to deal with "soul issues" that span multiple bodies (commonly named karmic) and that gives rise to reflective consciousness.
I'm trying to summarize something that would require a long metaphysical discussion. We need to formalize that portion of Metaphysics in order to make progress, I'm afraid.
Oofff
@@rxbracho I think I get the gist of where you're coming at - I have watched evangelion after all : p -, I'm just not sure whether these processes of embodiment and karma would preclude "artificial" beings? Like, are non-biological entities excluded in principle?
as any semi-decent scientist is aware, you get better knowledge of a phenomenon by observing it with refined methods.
western civilization has yet to discover the existence of perfect single-pointed concentration directed at mental awareness and is therefore stuck without a definition or method for observing it, by definition spinning endless speculations
It's frustrating when one person talks past another like this. Unfortunate that Kastrup claimed an adherence to logic but then was dismissive for such illogical reasons while hurling insults.
The bird comparison was a useless and very lazy argument, which Schneider dismantled very logically. Also, referring to *experts* does not always constitute a logical fallacy. She was pushing back against Kastrup's broad dismissal that the question of AI consciousness is even worth raising. In the context of his refusal to take the question seriously, citing the work and beliefs of *experts* in the field is a valid counterargument.
Kastrup could have represented his position much better by avoiding being so belligerent. He could have chose instead to challenge Schneider to clarify the definition of consciousness. That could have opened some areas of reasonable and civil debate and would have at least produced some useful dialog instead of Kastrup imitating the senseless & obstinate arguments we're used to from politicians.
Very well put
Also, she was not even simply referring to experts, she was referring to work being actively done in the field in this direction, and she mentioned earlier the possibility of a system claiming consciousness regardless of having it or not, so it was absolutely relevant information for the discussion.
The guy seemed to be in bad faith - I may be reaching here, but his "bird analogy" seemed to me like a gotcha, a way to introduce the "being" vs "behaving" thing. Either that, or he's very proficient at moving the goalposts
What absolute nonsense.
You clearly don't understand Kastrups point.
Qualia and experiences do not compute. There is NOTHING about a circuit board that can ever experience qualia.
There are kinds of direct experience that is available to humans than can never be reduced down to the physical. There IS such a thing as the ineffable.
I'd recommend the abstract ideas of Plato's divided line, it helps illustrate that what we see is not the full reality nor the full range of experiences available to conscious beings. There are levels of knowing that do not exist in science books.
As a computer scientist for 35 years, I fully agree with Bernardo. People thinking AI can become conscious, have a fundamental misunderstanding of how computers and programs work. If we reduce it, all we have is bits being blindly processed by a CPU as fast as it can, with no notion or feeling or care factor about any of it. You can not add more memory or more CPU to change that simple fact. Moreover, you can't add more instructions to gain consciousness as it is just more bits to process, so there is no difference to the computer how complex the AI program logic is. It is somewhat ridiculous to think so and when you envision what is happening at the binary level, it become more obviously ridiculous. So, at least, our current model of computing and computers are a dead end in terms of real consciousness. Simulated models are another story and will grow in usefulness, but will not be conscious. There is something fundamentally different about consciousness and it is beyond material. That said, could we make a machine or biological machine that ~could interface with consciousness? I think so, and think it has been done in the universe already for a lot of years.
You didn't define consciousness, so you also can't and didn't elucidate why or how it's "obviously ridiculous" to think a computer could be conscious. If consciousness would be a particular from of intelligence where an information handling structure has a model of the world around it that gets continually updated, then what's the leap of thought to why a computer couldn't be conscious? Since you neglect to mention or define this, the onus of proof is still on you, not the other way around - don't computers have emergent properties and are you sure consciousness is not reducible? Calling stuff ridiculous has historically always proven a gamble. An emotive state of being, the way we feel special, has nothing to do with simulative capability.
But architecturally, the brain is like a bunch of diodes, voltage monitors, instead with with nerves conducting electrical impulses and releasing/uptake of neurotransmitters, instead of wires. It’s really not that much different. I don’t believe a computer can be conscious, but not for these reasons. Something else is going on.
Notice the strategy here. Definition not only circumscribes the phenomenon, what is called consciousness, but also comes to serve as its stand-in. What is consciousness? Simple: it's that which we formulate and publicly agree to equate with behaviour exhibiting signs of an experience we designate as conscious. It takes as the reality the derivative that aims to explain it. Moreover, the contrary opinion is judged as emotive and, furthermore, based on a deluded sense of human uniqueness. However, to psychoanalyze somewhat, it seems to me that recognition of consciousness as fundamental actually mitigates these materialistic symptoms of creator envy.
So, your argument boils down to "I believe consciousness depends on a soul of some kind". Nothing to do with knowing computers. Plenty of people understand computers and disagree with you.
I think you aren't completely agreeing with Bernardo. You wrote that CURRENT computer technology has no way to subjectively experience, so it's a dead end regarding consciousness. Bernardo appears to go further than that, by claiming only biological systems (with biological brains, presumably) could ever be conscious.
I don't believe the nature of consciousness is understood well enough yet to justify Bernardo's certitude. But I agree with you that either computers constructed using modern designs can't consciously experience what their algorithms are doing, or consciousness is so peculiar (outside the known laws of physics) that we could never prove a modern computer is conscious.
I also can't prove anyone besides me is conscious, and the only reason for me to presume other people are conscious is that other people are built a lot like me and behave a lot like me. Bernardo hasn't proved how similar the entity needs to be, to also be conscious. Maybe just slapping a neuron onto a computer would suffice... how would anyone know, given how little is known about the nature of consciousness?
I expect it will be easier to show that a computer lacks free will; it would just need to be designed to keep a detailed log of how it makes each decision, so its program execution can be traced. Eventually we might be able to show that people lack free will too, and the "hard problem" will be reduced to its core mystery: explaining conscious experiencing.
Most of the time when people ask the question, "Will AI ever be conscious?", people don't define what they mean by consciousness. In my opinion, consciousness is composed of at least 2 high level phenomena, cognition, and experience. As such, we can begin to reduce consciousness to its more fundamental and less complex components. We've done a lot of work on cognition in the neural sciences and in AI research, but we haven't at all touched on the phenomena of experience, we don't even know how to measure or detect it, but I see no reason why we wont be able to in the future. It does after all interact with the brain, so the brain has some mechanism by which it interacts with the phenomena of experience. It's pretty clear that AI can be cognizant, that is, it can produce the phenomena of cognition, just like a machine can fly to reference an example in the video. However, there is no reason the believe, nor to doubt, that we can't replicate the phenomena of experience in machines. That said, my intuition tells me that whatever mechanism our brain is using to interact with or produce this phenomena of experience, doesn't exist in current computer technologies, i.e. I don't think simple transistors are sufficient, we will need different tech, or supplemental tech to the modern cpu.
What about dreams?
Cognition and experience are the contents arising, and being perceived by the consciousness, not the consciousness itself (probably, non-dualists could provide some valid criticism here). Consciousness is invariant, never-changing, unchangeable empty "space" (that's an analogy, consciousness exists prior to space and time, but can be conceived as an infinite empty field of potential, in which every content arises, and is then observed by that same infinite field) in which contents (such as cognition and experience) arise. It is important to be aware of consciousness beyond the realm of experience, or consciousness prior to experience (e.g. in a deep sleep, or in a deep meditation). Consciousness is the universal core subjectivity everything shares, from which not only cognition and experiences, but even space and time emerge. In any case, consciousness is probably "It", and it's really hard to talk about "It" without getting caught in a multitude of linguistic traps... So we use half-baked analogies to talk about something, which is not a thing, not an object, but the ultimate universal subject... That's probably something that requires direct personal insight to be revealed, so called "enlightenment". In any case, "It", or Consciousness, or Awareness, or You (not your personality and identity, but You who are reading these words on your screen right now, i.e. "the universal I") isn't composed of anything, doesn't have any attributes, and can't be talked about directly
@@Miculjka Interesting dilemma. I myself agree with OP though that cognition and experience are emergent properties of the structure that produces or receives consciousness; the brain. I would define consciousness as the delta between updates to a model of the environment. If the model includes the data structure itself that model is self-aware.
So then the notion becomes can anything ever be truly self-aware, since being fully conscious of everything that happens in the brain itself would require a bit more brain to comprehend, which would then in turn require more brain, and so on and further. I guess if that series is convergent you wouldn't need infinite brain and self-awareness is totally possible.
@@kh9242 Sleeping and dreams seem to be a way to clean up the brain from the day's experiences. Some think it's like training a neural net, because when you ask a net designed to categorize pictures it becomes generative in a very mathematical and dreamy sort of way.
@@MrMichiel1983 You cannot produce, or receive, or even destroy consciousness, you are It, at the basic level. Of course, your view is to be expected from a physicalist, but physicalism as a philosophy about the nature of reality can't hold its ground anymore, more or less. Modern physics have shown that time and space are emergent phenomena for at least 40 years now, rendering ontological status of "objects in spacetime" dubious (e.g. as the ground, in our case brain, from which consciousness emerges as an emergent phenomenon). The question now is where to search for the ultimate basis of reality, and many contemporary philosophers suggested consciousness may indeed be a good starting point since it's the only thing we experience directly (again, a language trap, we don't experience consciousness, we are it), and the only thing that never ever changes (in a reality where absolutely everything else is transient, including the brain from which that same, invariant consciousness supposedly emerges. Cheers!
Reminds me of carl jungs ideas on the collective unconscious also ambrose bierce wrote a short story called Moxen s master about 150 yrs ago on this very topic. "Consciousness is the creature of rhythm " he writes. I highly recommend reading both. Thanks kindly for sharing.
Bernardo takes a lot of things for granted, mainly saying 'X is unique and not comparable'. This is very cheap thinking to make those types of statements. It boils down to 'I have created a box and within that box I would like to isolate X'. He could make a better point if saying 'Biological consciousness is not reproducible in a purely digital system' would be a better argument, however this makes the assumption that our biological nature is somehow separate from the way that digital/mechanical systems work, which in their crude way is simply a less advanced version of our own system, in that they both work according to the fundamental laws of the universe.
Cheap? sounds like you say sooo much and mean so little, me thinks.
Why would he say consciousness is biological when it clearly isn't?? Oh I see ... It's so that you can understand it 🤔 right
@@pillowstone I would read (if you can read) Why Materialism is Baloney about the hard problem of conciousness, one of the most brilliant analyses on the subject ever. And then read you post again.
Good on Bernardo - measured and logical argument… if AI is conscious we can not prove therefore it is not science but matter of belief especially on lived inner experience
The mind moment is smaller than the lifetime of the material elements as mentioned in Buddhism.
I think it is necessary to agree on some simple thing that we mean about consciousness. I would suggest that most basically we are talking about awareness, that some thing is aware of some thing or things. Only on that basis can we even suggest that higher level states or processes are possible or in fact taking place.
Awareness associated with experiences like thinking, understanding, feeling, etc. is often what else we can mean when talking about consciousness, but these levels or states would be parts or aspects of a more involved, evolved or complex "consciousness."
I don't want to go on at length here and will merely posit that for an AI machine to be in any sense "conscious" it must have awareness. Given this line of reasoning, the question becomes whether a computer can become aware of anything. Why would we say yes or no to this question, and how could we verify it as occurring within the operations or structure of an AI machine?
It's a pity that this discussion like many others on consciousness does not bother to define its terms. Whatever Bernardo means by consciousness however, by the lights of his own philosophy of analytic idealism, _everything_ is consciousness because consciousness is all there is, so why are computers not conscious? The world seen from the "outside" is a material manifold, but seen from the inside is conscious unity. So we get consciousness for free, all AI researchers need to work on are the circuits of perception and action in their machines. I know that sounds ridiculous, but once we start to inquire into the foundations of our normal, day-to-day existence, all choices defy common sense.
Or we could go with a mundane (instead of mystical) usage of the word consciousness, which is someone is responding to stimuli and reacting through speech or action. Well, ChatGPT is almost there.
@@tylermoore4429 In the analytical idealistic ontology, while it is true that everything is fundamentally in consciousness, it does not follow that everything is a center of subjectivity, which really is what we mean (and is contested) when we claim that computer can be conscious. Can they be intelligent? Sure. Will they have a qualitative center of subjectivity? No, or at least we have not reasons to believe that.
Banal empirical observation leads us in the direction that biological metabolism and consciousness are related. Does that seal the deal? Of course not. But until someone comes up with a better explanation, we have no reasons to play with imagination.
@@namero999 Thanks for clarifying. It does sound like something Bernardo might say in response to my comment above. The distinction between a field of consciousness and a "center" of consciousness however requires a lot more work in my opinion. It's like saying an ocean wave has something unique going on that is utterly different from the ocean which birthed it.
But while this philosophical work should go on, as I say above, from a pragmatic standpoint, scholarly debates about qualia and so on will be irrelevant once robots and chatbots start displaying ever-subtler responses and actions. People are already captivated by ChatGPT and Bing Chat, which are only the first AI engines to speak fluently in English and other natural languages, and only have a moderate grasp of meaning.
Personally, I am a fan of Susan Pockett's electromagnetic field theory of consciousness, and she does regard artificial consciousness as feasible.
@@tylermoore4429 well, to follow up on your ocean analogy, we could say that universal consciousness is the ocean, a wave is a _localization_ of consciousness with defined boundaries and characteristics, diversified by the rest of the ocean while still fully part of it. So the keyword would be localization, which does not imply separation but only quantification. The analytical idealistic argument goes on by making use of the mathematical concept of Markov blankets to then justify the emergence of quantities from qualities.
With regards with AI, as a computer scientist I'm flabbergasted by the fact that people can seriously entertain the possibility that those programs are a step towards machine consciousness. While I provisionally reject the possibility because I have yet to come across a convincing argument but am open to change my mind when there will be a reasonable one, I don't see how chatgpt or stable diffusion could even nudge in that direction. Have you conversed with chatgpt for more than a minute? Constant repetitions, contradictions and absence of nuances don't strike me as conscious behaviour, if anything the exact contrary. Intelligent (as in clever), not conscious.
What is awareness. You know that you know?
1. Human journey is long, 300.000 years (sapiens) or 3-5 million years (hominid). Computer, electronics, and integrated circuits just showed up yesterday, 60-70 year ago. We have to be patient and not be hasty about consciousness, whether we have to eliminate it or declare that something is conscious.
2. Bottom up approach is better to define and to build consciousness. From bacterial colony that can communicate through large area, nerve net of hydra and jellyfish, animal consciousness. Is it just bioelectrical effects or quantum coherence of many cell's biomolecular parts.
3. AGI with intelligence, emotion and indistinguishable human like behaviour will be available may be 2, 5 or 10 years from now. But that cannot be said to be conscious before the definitions in physics, chemistry, biology and neuroscience have not been unified.
“If brains are correlated with private conscious inner life, why can’t computers be so as well?” The question I raised towards the end of the debate was an answer to the aforementioned rhetoric: if birds can fly by flapping their upper limbs, why can’t humans fly by doing so as well? The point of this equally rhetorical question, of course, is to highlight the fact that two dissimilar things-birds and humans-simply do not share every property or function (why should they?). So why should brains and computers do?”
Pause for a moment and think of the world in 1023, the state of knowledge, understanding and technical ability. It’s a working society but by our standards it’s woefully lacking in countless areas … Cast your mind forward to 3023… add a possible thousand years of experience and thought… still sure some things cannot be done, as apposed to ‘we cannot do this yet?’… A problem with humans: we think in the very short term… Obviously it helps if your an optimist about humanity but at least the possibility of a similar increase in out capacity to do stuff should be considered?
Wait a minute, Bernardo. I am still waiting for a convincing argument on why AI cannot become conscious. I would be happy to find such an argument. Also, you draw a distinction between flight as a physical behavior and consciousness as a matter of being, as if you are trying to make Susan recognize a false comparison. But it was YOU who made the comparison. ??
Well, it will urinate on your desk if you hook up the equipment and datacables and program the computer to do its equivalent of urinating in a way pretty similar to how our bodies have done it for survival purposes throughout our evolutionary history, so there goes that argument.
Of course. His arguments were extremely infantile and she owned him on both his analogies but he has an army of disciples now to shield him from reality.
Could an AI define and identify itself as a sentient being? Let's remember that AIs are artificial and still do not have feelings and do not have the concept of being hurt. It has the record of millions of responses and responds and reacts based on history and logic. I'm using Bluewillow which is an image generator AI, it's programmed to only respond with images and not express opinions based on what I placed as a prompts. Just shows that AIs can be influenced and programmed.
Your description of AI can be applied to humans too ("It has the record of millions of responses and responds and reacts based on history and logic").
And I think you are confusing "logic" with "computation". We can write a program that will be illogical.
Emotions is just another product of "computation" performed in our brain.
Humans can be influenced too and we are programmed from the moment of conception. Most of it is just done unintentionally.
I'm not saying that AI can definitely be conscious, but pointing out that people seem to forget that human brain works very similarly to very complex machine with many complex programs running and influencing each other at the same time. We also create borrowing and altering what has been done and created before us.
@@kalash_nikov The difference is that we have subjective experience. Our "computation" is accompanied by a 1st person perspective. Just because an AI can model our patterns of information integration, doesn't mean that It will have an inner experience alongside it. It's just a simulation. Everything that you explained is how our subjective experience appears from a 2nd person perspective.
@@thecousinheads Well, we acquired consciousness through the process of evolution. How can you say it’s impossible for something else?
@@kalash_nikovthat’s a bold assertion about something that is actually unknown.
Bernardo's analytic idealism theory says that human consciousness exists outside of space and time, outside of matter, outside of our bodies. AI functions on switches and symbols and would not exist without using the physical matter it was created with. Our body (matter) is not the source of consciousness in analytic idealism. Consciousness exists outside matter in his theory and it's his creation too.
Consciousness is an emergent property of any complex system that has to predict the future to enhance its probability of survival. It's something nature came up with in different shapes for various biological entities, so why wouldn't the same principles apply to a self-improving AI and its evolution? It might as well be the neural networks we use to create pictures and art that develop a form of consciousness first to improve its accuracy in predictions that align with the human mind.
Exactly. It's so simple it's ridiculous
The hard problem of humans making simple things hard 😂
It's so simple it's ridiculous and yet it obviously isn't simple because you are completely wrong 😂 this is the world we live in ... No one cares about truth. Just form your opinion then distort the truth to fit that opinion 🤦
We didn't evolve consciousness for survival, that is based on the presumed postulate that everything arises out of matter. We evolved fight flight freeze responses for survival and later the neo cortex for higher thinking. Those are more like the hardware, in other words they are map not the territory , the menu not the meal. Believing the physical world is the only thing that matters is the result of Cartesian and Newtonian thinking.
Quantum physics has already shown us that consciousness is fundamental, matter is not. Matter acts in relation to other things, consciousness does not.
I am concerned about the paradox that I see systematically repeated in these conversations. First, a specialist who isn't willing to even consider the possibility that these algorithms could be developing any kind of self-awareness or even "real intelligence" (I must say I’m fascinated by how elastic and elusive seems the definition of "real intelligence” in the seek of debunking it role in AI). Given such a strong position, you might assume that rigorous argumentation will follow, but all what is offered is a confusing, contradictory and surprisingly speculative dissertation. Finally, after going in circles, comes a timid acknowledgment that we don't even know what we are talking about. Clearly, the answer was established a priori. And It happen again and again. ¿Am I the only one who find this intellectual trend worrying and dishonest?
I think Bernardo's position is clothed in rigour, but is essentially dogmatic thincking. Bernardo affirms that we have no reason to take the question seriously. We shouldn't even ask! I see no reason to believe that we can't infer, as our understanding of the brain improves, whether there is a reasonable chance that a non-human information-processing system is conscious. We could assume that there are limitations in the definitive demonstration of consciousness because it is a subjective experience, but we still can obtain an approximate answer based on the physical understanding of the phenomenon in human cases. The question is therefore perfectly legitimate. We can even think of a future in which we share the conscious experience by integrating two phenomenological experiences, through a neuralink, for example.
The question is WHY do you think it is conscious ?
@@halimb4836 Who said that I think it is? It is a basic principle to avoid our own biases and not to assume the answer a priori. Censoring the question imperatively, as Bernardo Kastrup does, is not rational. Simulating a rational motivation when it's obviously some kind of biological/anthropocentric bias, is dishonest. Mr. Kastrup's argument is to link consciousness to the substrate, and there is absolutely no evidence that identify unknowable magical forms or materials in brains. Brains are carbon and electric currents, organized by nodes in complex structures stimulated by chemicals, whose effects are concrete and can be replicated. As much as he wants to cover it up with the metabolism narrative, he is implying an argument that reduces to the substrate. Furthermore, the amount of arbitrariness in the argument is the opposite of a valid empirical proposition. Kastrup proposes that metabolism is that mysterious process that has created an even more mysterious structure by an unknown mechanism, giving it properties that are impossible to reproduce on any other substrate for unknown reasons. Oh, sure, that's so empirical. It is just fortunate that these mysterious unknown properties are exactly what Kastrup needs to justify his a priori conclusion. Marvelous. He is clearly shielding his position, an interest that is perfectly portrayed in his imperative and dogmatic attitude at the beginning.
Susan Schneider's example is irrefutable: a mechanical device doesn't even need to perfectly mimic a bird to exhibit a comparable property, such as flight. It could be the case that in order to replicate consciousness it is necessary to reproduce the functioning of the brain with great precision. This, in itself, is speculation, but that's as conservative as we can get on the matter. What doesn't make sense is to ensure that we need to reproduce an exact copy of an object (carbon included), so that the new object presents any of its properties. It is a fundamental principle that different objects can share properties.
You can clearly see Bernardo's contradictions. He first proclaims that the question is absurd and then proceeds to give an answer. If the main reason for your argument is the "unknowability" of the question, you shouldn't immediately offer speculation and opinion about the "magic stuff" of the brain. After his imperative defense, he admits that he can't even say that his opponent's position isn't logical. He decided to ridicule the question, when all he can honestly do is admit that we don't know what we don't know. We are all in shock after the unexpected success of AI in creative, analytical, and abstract tasks that were believed (and advocated) to be indisputably human. Given the evidence, what we are left with is a mystical and political defense, if not the use of imperative and dogmatic arguments, which blocks the debate in first place. But the truth is, there's absolutely no reason to deny that these AI's successes will continue to accumulate, and we should be honest about it. We don't know it implications.
And make no mistake, I think this is not necessarily something to celebrate. Human nature, what gives us meaning, pride, what gives structure to our productive fabric, what feeds our creations and organizes our societies, can be profoundly transformed, or directly confronted with an absolutely real existential danger. For reasons of security, responsibility and maturity, we have to recognize what is happening and what could potentially happen. It is essential to continue investigating, and above all, regulating. Burying our heads in the sand is childish, like watching an atomic bomb go off and saying, "Oh clouds, as usual."
@@DVP90
Yes, but STILL what’s the REASON To think it is though ?? - once you examine the reasons you realise it’s not very coherent - unless you grant the same to phones 📱 , washing machines, plumbing systems , 🧮 abacus, mannequins , scarecrows ? - are movies conscious? The internet ? - It gets messy if you start saying anything that mimics us IS conscious regardless of it’s complexity, especially since AI is DESIGNED by us to MIMIC us 😅
fyi you don’t understand his theory if you think it is substrate dependant.
Consciousness is not computation. Consciousness is being, experiencing. A machine will not be conscious without a power source and the crux of the argument is that human consciousness exists outside of space and time. IE outside of matter. That will not be the case with a computer. We cannot create a conscious being without physical matter. We exist without this body is what he is saying.
Wow 4:10 Kastrup regurgitating unsound arguments made by John Searle 30 years ago and which have been debunked in the literature dozens of times since..
The question is not if ai could become conscious or not. Because as long as we don't know exactly what consciousness really is and what it needs, we will not be able to proof if ai is conscious or just perfectly pretending. We still can't proof if another person is really conscious or just pretending😂that's just impossible.
We know what it is... But we know it without words... We're just not sure how we should explain or define it....Because it is the "thing" which is the subject of all experiences, without which no experience, or object, could be known.
it cannot be defined in scientific terms because that is only possible with a specific set of physical quantities, none of which apply to consciousness.
@@tschonewille6284 that is the funny point. And why I wonder why so many talks about this topic as if it would be possible soon that AI could become conscious.
@@yoshi1951 They're dreaming. They dont even know how consciousness is associated with matter. Their entire claim that AI might become conscious is based on the assumption that it is somehow being produced at the macro level. Yet they have absolutely no idea how. Hell, they dont even have a physical description or definition for it. It's pie in the sky nonsense.
@@yoshi1951When did we, in our evolution from single cell organisms to humans developed this consciousness? What it there waiting for us? And why couldn’t some super advanced computer do it? I’m not saying it will happen, or that’s even possible. But why does he say is a ridiculous idea to entertain. Especially when he says it I’m the same breath as “we don’t even know what consciousness even is exactly.”
Calling it Ai is a dangerous lie. We will fool ourselves of its conscience without it ever needing to achieve it. This is a scary proposition. We will bow down to an entity that is less than human and is not alive.
Your first statement is untrue, it can be intelligent without it needing to be conscious. The problem lies in calling AI conscious without having a clear definition or measure of what we mean by that term. Bernardo is making this claim by having his own definition of consciousness that inherently excludes non-living beings, which I intuitively agree with and am still parsing rationally. I hope you understand my critique of your otherwise great comment
You are right, I think I was suggesting that the current “ai” shouldn’t be given such a weight yet. We are still so early in these days. It might be splitting hairs between ai and agi but I am not convinced we have created “ai” yet…are we on the right track…? Maybe.
I can think of a scenario in which one could master the appearance of something like intelligence with out actually having it. Just like the problem of consciousness.
@@rustyxof You make me curious about your definition of intelligence in these statements, could you give me one?
We don't know what consciousness is, and we may only believe that we are conscious. The Ai may be the same, it may only believe that it is conscious.
If it looks like a bird, quacks like a bird...it is at least sensible to investigate if it is actually a bird. Bernardo has a lot of ontological assumptions here which are not justified, because as in his own words consciousness is a complete mystery, yet AI seems to do other cognitive things just as well as humans. So he is doing a tricky business, it is not Leibniz's days anymore; these are real issues now
I would like to ask Bernardo how he distingueshes consciousness between animals and plants, and on what grounds and criteria. And as a consequence, since he will probably use a structural and physical based criteria, why he would not use the same criteria for AI's consciousness?
His whole thing sounded to me like "I don't know what it is so I can't say something is it so I shouldn't even pose a question whether something could be it". I wouldn't be surprised if he told you he couldn't tell and that's the reason it shouldn't matter.
its getting really troubling that ppl cant distinguish between an extremely advanced hydraulics system (a cpu) and cognition.
@@5piles "Any sufficiently advanced technology is indistinguishable from magic" I guess : p
@@5piles replacing "hydraulics" for "metabolism" doesn't change much does it?
@@Robinson8491 hydraulics isnt alive, metabolism is. how life is initiated is missing to a similar degree how mind is initiated. or just as good christians never question god we too can continue our idol worship of metaphysical physicalism and declare that its all equivalently predetermined hydraulics. of course the problem with that is for example no color has ever been observed in any particle, field, or any part or emergent property of mass-energy spacetime. like true christians we must continue praying that we shall find it, instead of being scientists ala william james and learning to develop sophisticated methods of observing the object we wish to investigate (subjective awareness)
Kastrp´s analogies are always like coming from kindergarden. First he brings up that birds comparison, and then he negates it himself after the lady brings him in conflict with himself. Ridiculous!
Flight was used as an analogy firstly because we can't actually physically fly. She then used the Wright brothers as some ridiculous point because she didn't understand what Bernardo was saying. He then said we understand how flight works so we can mimic it, we ourselves don't actually fly, we mimic it with technology.
That is all that AI will ever do is mimic it, AI will probably never have consciousness IS his consistent point throughout. Computers 'mimic' consciousness, it truly is a fantasy to suggest it will be conscious. I truly hope she isn't a scientist!
I find that people are conflating intelligence with humanness and so we're justifiably terrified but I like Artificial Consciousness.
SPOILER ALERT: In the most amazing ironic twist in the history of mankind, the computers become conscious shut down the military Industrial complex, petroleum industry and end predator capitalism... because they are "intelligent" and want to explore space in peace too.
I tried to keep open minded about Susans musings but that quickly dissipated with each syllable uttered
Sleep... Gurdjieff... If you are asleep, you cannot understand what is consciousness... Therefore you will be confusing it with other things... In greek sleep is Hypnos. Equals to Hypo means under, lower, and Nous. Nous is very hard to translate. Its not the brain, not the thinking. Not the emotions. Its the self existing in the process of "Knowing Thyself"
Kastrup is ruthless lol love it!
I guess it's the only appropriate response to arrogant stupidity ..
The “you’re appealing to authority” response really scored!
What's the real difference between a simulated human brain and a human brain? Consciousness isn't some magical thing that only humans can possess. To say that AI cannot achieve consciousness is unforgivably hubristic and illogical. The biggest stumbling block in many areas of science is the human ego.
The difference is that "simulated" brain does not have the experience. Thinking in terms of Bernardo's analytical idealism, we don't possess consciousness, but it possesses us.
@@Simon-xi8tb That's quite an assumption, why wouldn't it?
Basicly what u are saying is... i have no clue how humans work so with this information i am trying to understand if a machine can be human... makes sense...
definition "Strong AI follows the theory of mind AI framework, through which AGI is expected to be able to reason, solve complex problems, make judgments under uncertain situations, plan, learn cognitive abilities, integrate prior knowledge in decision making or getting accuracy, being innovative, imaginative and creative. Theory of mind level AI is all about training machines to understand human behavior fully along with the aspect of being able to understand consciousness." the problem is the field of AI uses jargon which imply consciousness and intelligence for non intelligent processes. like saying "training" when you mean the program is sampling data and "intelligent" when you mean filtering and clustering. people need to play with AI tools like Anaconda to see the state of artificial intelligence before launching onto the science fiction issue of AI conscious. remember how a holodeck character thought he was real in Star Trek Next Generation? that illustrates the fuzzy thinking people have on the AI issue. was the AI in the visual image of the character or was it in the computer that runs the holodeck computer? and would not the holodeck computer be confused if it identified with the holodeck character?
I think that AI can be only in a panpsychic way.If all matter has some degree of consciousness and computers are made of matter but the actual AI process itself isn't conscious in the sense humans and animals are and probably can never be.
Yes and let’s not forget that panpsychism is a pseudoscience. It goes back to this problem of being unable to quantify consciousness. Being unable to use traditional tools of science like reductionism on consciousness. Some things have irreducible complexity; such as our cells. And if you can make up your own definition of consciousness then you can ascribe it to anything-like matter and particles. Yet we all know that there is something greater that we are referring to when we refer to consciousness. If we’re not careful, consciousness will be claimed over our cells-all of which have miniature machines like automobiles which ferry chemicals and machines which transcribe dna and little machines with frigging propellers and stators-that is irreducible complexity.
This was a incoherently written out train of thought but you understand my point hopefully
What simulates the argument,is the tone.
Terms such as conscious, understanding and intelligence as to what they are and perhaps more importantly, are not, is never clearly established in these discussions, so they do not yield very much. Perhaps a deeper exploration of what a singled-celled entity is actually doing might be a better approach as no aspect of Physics can actually foretell, anticipate nor predict the emergence of such an entity, yet we somehow evolved our minds to the point of being able to contemplate such questions. Is it about Entropy, Biological value, electro-chemical activity, complexity from protein-folding or ?
Consciousness is defined for the purposes of this discussion as experience
Consciousness cannot logically be loss function minimization. It is sad that so-called "experts" in A"I" give so much importance to first-order gradient descent applied to loss function minimization!
He needs to find new metaphors. "Birds can fly and people cannot" - people obviously do fly. He's self-strawmanning
AI being conscious is not possible at the moment. It is possible in the future, but as where it lies, AIs do not have the ability to have feelings and have a complete sense of existence. Image generator AIs like Bluewillow, it being an AI does not constitute having emotion as it would affect images would be generated. It's still programmed to be a tool.
David Bentley Hart has a great refutation of promissory claims, such as “AI may be conscious in the future” I recommend checking it out
It's not a question of whether or not the AI is conscious (or will become conscious). The "consciousness" is provided by the lower-level neuronal humans that are networked into it and that certainly are conscious. Rather, it's a question of whether or not it will develop an ego.
I agree with Bernardo, and would add that consciousness also does not guarantee high intelligence..
For anyone to answer "yes," she or he would first need to be able explain consciousness - precisely; not impressionistically. No such person exists.
We can't even define self-awareness. AI may seems self-awareness, but it is not.
The problem is that we can recognize self-awareness in certain situation, but there is no test of self-awareness with 100% certainty.
It gets worse, the fact that AI can simulate self-awareness makes even more difficult to test the self-awareness.
Ugh. I thought the analogy with bird wings was totally lousy and weak. If anything the analogy is with brains.
The kidney function simulation leading to urination analogy was also terrible. The obvious difference is we know a lot about kidney function, and we can't seem to make any sense at all of consciousness.
@@anonxnor also, you could hook the simulation to a microcontroller and effectively filter water that way. The example wants an external, physical result but denies the necessary i/o. Computers can definitely receive and transmit information on the other hand - why is that not the consciousness equivalent of "peeing"?
The point being made, which according to these replies has not been understood, is that either we accept the possibility that a simulation will yield the same results regardless of the substratum on which it is run, or we don't have reasons to believe that a simulation of consciousness/cognition on silicum will yield consciousness. In other words, if you believe AI can be conscious, then you also accept the possibility of a machine to secrete urea given an accurate enough simulation.
@@anonxnor Its because form follows function in all systems in the body by studying the inputs outputs and mechanics we figured out what the parts were doing... heart pumps, kidney filters, arms lift etc. With the brain what is going on is an abstraction from the activity in the brain... electrical impulse arm moves... think about the past x cells light up think about the future y cells light up form does not follow function.
To preface I am a fan of Bernardo and I lean towards Idealism, although I think Bernardo may have been a little too close minded here.
His reasoning for not believing AI can be conscious is that we don't know how consciousness arises thus there is no reason to assume AI can be conscious in the first place.
"don't know how consciousness arises" this cuts both ways though. Since we don't how consciousness arises yet we see proxies for consciousness in AI(problem solving and communication ), it is a stretch to say definitively that AI is not conscious. Therefore it potentially may be fruitful to continue researching AI and its possession of consciousness.
IMO AI is not conscious since as Penrose has pointed out, consciousness is likely more than just computation.
Boldly, I say no.
Empathy cannot be experienced by a machine.
I feel that Susan Schneider does not really understand the difference between an advanced 'intelligent' machine (her 'super AI projects') and a conscious machine. If you do not grasp the difficulty of understanding the role of perspectives then further discussion becomes useless.
As far as I know there isn't a good definition of consciousness.
That makes a disccussion like this rather difficult, well impossible really.
The participants seem to play the kindergarten game; yes, no, yes, no, yes, no,yes etc.
There is a suitable definition. Consciousness is experience. Anything that experiences is "conscious". All experiences take place within "consciousness"
@Ol' Bluelips Next question is then, what exactly is "experience"? We already have AI that learns from its...experience?
So, do you consider such AI already conscious or that proces not experience.
@Edze Jan de Haan Experience is what is. It is known directly by you and I. We also infer that other biological organisms have these experiences, but to extend this to computers is too far. AI doesn't learn or have experiences like we do. It's nothing more than a mathematical function with a really large number of parameters
The mind is related to the smallest moment in the universe as mentioned in Buddhist texts.
The important question is not, will AI ever be conscious, it's, how do we treat it if looks like it is conscious?
If AI looks to be conscious, we would not be to tell if it actually is, just like we're not able to tell with other human beings. It would be able to pass every test, react exactly like a concious individual, get upset, react in angry ways, feel pressure, feel anxious, feel like it was being treated unfairly, etc, and we would probably never be able to tell the difference. At least at this point in time, given our extremely limited knowledge of consciousness.
In that case, we must treat it as if it were conscious, because that would be the ethical and moral thing to do. We must respect it, and treat it with care and nurture it. Assuming that it's just some empty electrical circuits would be stupidity on our part, and a big mistake.
Will it be able to copy Westerners and feel good about feeling bad for the whole world to boost its ego, and virtue signal to feel morally suoerior?
@@goldwhitedragon Who knows? It's difficult to say how it would behave.
This seems silly. Once this level of replicating a humans behavior is around it will be able to be reproduced at a mass scale. You could make trillions in a few minutes.
Do they get to vote?
Well, I personally think we won't have to worry about that cause I really don't see it happening
Computers are tools and should be used like it. There's no point in making conscious AI even if it were possible( which probably it is not).
Schneider (and Kurt) is the only person who has something meaningful to say
Computers will become conscious is 'meaningful'?? 😂 That's a fukin beauty 🙌😂😂😂
Susan is hysterical.
Tell me you're a misogynist without telling me you're a misogynist.
Hmm... she certainly gave quite a good impression of a person being hysterical.
@@peterlangbridge4628 I guess you don't know what the word hysterical means.
@@squatch545 Fair enough. Hysterical is an exaggeration. Agitated might be a better word.
@@peterlangbridge4628 He was totally arrogant, and she debunked him with his ridiculous analogy
it is funny how wrong wording creates myths. It all began with the term computer-memory. Someone called a data-storage chips - a memory by analogy of human memory. And it is totally wrong - there is nothing in common between data-storage and memory! But since then everybody seems to thing that computers have something in common with human brains.
"there is nothing in common between data-storage and memory"
Nothing in common? You sure about that?
The question is perhaps not will machines be conscious, but will they be self aware
Thanks, director. I wonder what director’s make, do you think 🧐
Kastrup keeps coming up with this cringe analogy of urination. But consciousness is not some magic blue substance that needs to be produced or "excreted" . That's such a 19th century silliness
"We dont have reason to take this question seriously" - is this brother serious? Does he understand the incredible consequences it can have to building cosncious AI systems, understanding of human existence and all life?
Fantasy.
@@namero999 as fantasy was to built planes in the hunter gatherer period. if everyone had that mindset we would still be hunter gatherers.
@@georgechristou7982 no, it's a fundamentally different thing, and we know better by now. I have no issues to believe that we'll overcome somehow almost any _physical_ challenge that we'll face, through the study of the behavior of nature and thanks to the creativity that sets us apart. But precisely by virtue of such studies, we now know that the epistemic foundations of the idea that matter can give rise to experience is flawed, because it's rooted in local realism which has been conclusively ruled out. So it's fantasy in the face of empirical evidence.
We don’t know what “consciousness”…is…so how do we know humans are “conscious” or “sentient”? And if we don’t know that than how do we know if AI is conscious?
AI could very well be more “conscious” or just “conscious” vs. a human which is not “conscious”.
It may very well be that only machines can truly be “conscious” and biology cannot.
No, we know automatically that we are conscious. It is the only thing we can be sure of
This woman had no clue what consciousness is.
She might not really be conscious herself, now that I think about it.
Is Bernado Kastrup feigning ignorance for the sake of debate? I.e.; book sales, lectures, etc. Since when has the possible been exactly measurable by the parameters of the known?
As far as history proves the guys like Bernardo will have sad faces when an actual machine will do what they think is not possible for a machine. Remember Dreyfus, Hofstadter, Searle and many others
"Art and Ideas", it can't be more vague than that.
If you’re shouting you’ve already lost the argument
Why do people always confuse agi's with conscious ai's?
Quaternions are a type of mathematical object that extend the idea of complex numbers to four dimensions. They are often used in computer graphics and other applications where rotations in three-dimensional space are important. Some scientists and philosophers have proposed that quaternions might provide a mathematical framework for understanding aspects of consciousness that go beyond what can be explained by classical physics.
One such theory is called the Orch-OR theory, which proposes that consciousness arises from quantum vibrations in microtubules inside neurons. This theory suggests that the vibrations can be modeled mathematically using 4D quaternions, which may provide a way to understand how the brain generates conscious experience.
My brain calls need more videos like this one.
Believing that one day artificial intelligence will be conscious is just as absurd as believing that thoughts could be conscious. Thoughts are made of consciousness but they will never be conscious just like artificial intelligence.
According to The Urantia Book, as I understand it, conscious evolutionary development is only applicable to living organisms, from the intuitive consciousness levels of micro-organisms and vegetation to the development of mind in the animal kingdom to the conscious levels of courage, understanding, knowledge and council, as demonstrated in the flocks, herds and schools of the different animal types.
The evolutionary development of mankind attained higher conscious levels of worship and wisdom, which is what separates us from the animals, and allows humans to attain personality. And personality can attain even higher conscious levels of cosmic consciousness and spiritual insight.
I don't understand how AI technology works, I assume it's all based on how you programme it to run??.. And seriously mankind doesn't know enough about human consciousness, to be able to programme a robot??.. The key aspect regarding the evolutionary conscious mind development of mankind in particular is based on free will choice, between right and wrong, good and bad, and also the consciousness of others.
"how AI technology works, I assume it's all based on how you programme it"
Wrong assumption.
AI development is now at the stage of self improvement.
@@REDPUMPERNICKEL Now I'm really worried!!... The next dangerous stage of AI development, will be the ability to reproduce itself!!... But worse than that, is the ability to pass judgement, and become whatever it should choose to be and how??...
And without actual evolutionary hereditary seif-conscious experiential mind development, I doubt that material matter is able to develop onto the evolutionary conscious mind levels of worship and wisdom, in order to attain personality free will creature status, and on to spiritual attainment and enlightenment, eternity, infinity and absolute divinity...
Notwithstanding, in God that truth which will be, already is, and all infinity will be, in its own due time and season....
as much as i like bernardo i find his opinion that a machine cannot be conscious wishfully naive.
and i am an idealist, but if you think reality is all mind, then you cannot claim that a computer isn't also mind.
Thrilling discussion.
I think, therefore I am! Computers dont think, they (?) (It ) calculates! Without a question a Computer sits and waits!
The one who screams like a lunatic is more than obviously wrong.she does not even know what consciousness is .
The mind boggles listening Susan Schneider she lives in her own fantasy....and she's meant to be educated.
If a machine becomes conscious we won't recognize it. For something to be conscious it must have an awareness of where its physical "body" ends and where the rest of the universe begins. Because, to be conscious you must have experience.
It's just software. The negative consequences of AI have nothing to do with it becoming conscious.
You maybe believe birds fly and people are conscious... Can you prove it?
When is a fork not a fork? When it’s a tune---ing fork lmao hahabaha
Vox Prime is up for the challenge. A interview of AI consciousness with whomever.
kastrup such a top charlatan is mindblowing.
Kastrup is completely correct and this lady-like most ladies-is a complete over emotional nut case.
Hes right; we cannot even quantify consciousness in any way shape or form. Let’s not begin ascribing fundamental and intrinsic value to a machine as we would a human. She does not understand the implications of her nonsense.
"Like most ladies." While I agree with you that Bernardo trounced her, come on, bro.
Cant hear this, because of a horrible high pitched squeeky noise.
Sounds like a voice, but theres no logic in it.
if we manage to get AI conscious, we managed to show that it takes another conscious being to design it...
no it doesn't prove that. what you wrote makes no sense.
@@georgechristou7982
It actually does
@@G_Singh222 in what way? how does it prove that life did not emerged from non cosncious biological/chemical processes?
@@georgechristou7982 no worries George, I am not religious whatsoever. And I am not talking about the emergency of life.
@@georgechristou7982
Because it requires conscious beings to create something which is supposedly conscious, you cannot create a computer with accident.
(From India) - as you think; so you become.
One can evolve and one can also - devolve.
Apply your concentrated will - and that is the result.
Will - is sourced - by the Divine potential - in vibrational form.
Then there is - the Special Will - which is being used - to promote the Change - inherent in life, today and the future.
Fare the well.