Where do ye think that subjective experiences, consciousness and in Ancient Greek, ‘qualia’, come from? There are many hypothesis by religion and neuroscience. So far, no answer has been found. One might even say, the more we know about that, the less we understand.
At least, it shows how far the world is from creating a true human-like AI.
I have to admit, I got that question after somebody sent me pictures of the “NPC” meme.
Given the existence of unconscious perception (“blindsight”), it stands to reason that there is a special process that is applied to information present in the mind to make it conscious. Consciousness seems to be a holistic phenomenon of the brain. Information that is encoded to be processed by the holistic consciousness (holistically encoded) is that what we call qualia. We just don’t know how this code looks like. The code behind qualia may be simple, but may have eluded human ingenuity and intuition so far. Unless someone already secretly cracked it and hasn’t published the results, yet.
An important question is whether we will crack the qualia code before creating general artificial intelligence, or afterwards. That also depends on how important qualia is for the functioning of generally intelligent systems. If it wasn’t important, then why did nature create something that looks as wasteful as holistic encoding?
It could also be the case that qualia are necessary for neuromorphic AI, but that there are other forms of AI that don’t have consciousness and work just as well or even better. Imagine an AI that could tell you exactly how it works and what it’s thinking without having the slightest hint of any subjective experience.
Because without it, everything else would be pointless, since literally, nobody would notice.
That depends on how one defines AI. Some say “strong AI” is consciousness and “weak AI” is not. For me, a real AI has subjective experience of er own. E perceives, feels and thinks akin to a biological person. A fake AI, a program that just passes the Turing test and is indistinguishable from a person for somebody who interacts with it, but has no personality of its own, is a virtual intelligence (VI).
Der Panpsychismus (besser die Theorie der integrierten Information) lässt sich natürlich unter verschiedenen Blickwinkeln betrachten. Der von mir eingestellte Link beinhaltet, neben einer guten Zusammenfassung des Themas, den Blickwinkel des Theologen. Einen kurzen Betrag dieser Art aus dem Blickwinkel eines „Materialisten“ bzw. eines „neutralen“ Autors hatte ich so schnell nicht zur Hand. Daher vielleicht der Eindruck, dass es sich hier um eine Spielart des Animismus handelt.
P.S. In einem deutschsprachigen Forum sollten wir uns bemühen unsere Gedanken in deutscher Sprache auszudrücken (sofern die deutsche Sprache beherrscht wird).
What kind of reasoning is that? It feels teleologic. Nature did something, because there was a higher aim for nature without which everything would have been pointless.
That’s not how science imagines nature to work. It’s rather how an intelligent designer of this world would operate. But even if we accept simulators as intelligent designers, the needs to be a first emergence of consciousness.
What if it is possible for some kind of intelligence to not be conscious, but realize the teleological necessity for consciousness, and therefore creates conscious beings? That feels improbable, but on the other hand I can’t disprove this idea off the cuff.
Why do you suppose that understanding requires consciousness? Understanding may come from the combinations of unconscious insights, which could be defined as surprising (improbable) facts found out about the world. An AI may be able to collect facts about the world unconsciously and combine them, thus eventually getting some unconscious realizations.
The core of consciousness is qualia. Without subjective experience entities are not only unconscious but zombies. So i think we both differ in the definition of consciousness. While you chose the dichotomy “consciousness-unconsciousness” i prefer “qualia-zombie” which is equivalent to “alive-dead”. I am aware that it is unclear, what the unconscious mind really is. When somebody is unconscious while he undergoes surgery, we hope that he will not experience anything but we know more: People tell about dreams and some even seem to remember fragments about what happened to them while under anaesthesia. So are there really no sensations while being unconscious? I believe the unconscious mind has the function to reduce experiences/qualia and helps to forget experiences. But to forget something, doesn´t mean that there where no sensations at all, even if it is possible to forget something instantaneously. Many people often can´t remember what they dreamt. The subjective idea might be, that it was like being unconscious while they slept, but a sleep laboratory can prove, that they experienced a dream. In my opinion being unconscious means being deprived of sensations but not to zero, because zero qualia means death. If AI will be able to get some unconscious realizations, it will have qualia, no matter how minor.
The expression “realize” usually refers to a special kind of qualia somebody experiences. We could say that a thermometer realizes the temperature and a thermostat even reacts to the realization of the temperature to help us save heating energy. But nobody believes that this “realization” has something to do with qualia and the sensations humans experience when they realize something. The real interesting question will be, if AI will remain something like a thermostat, even if it is a complex one, or if some quantity of complexity will be enough for the transition to become an entity with qualia and therefore a conscious entity.
That sounds like a strange function. Why would it be necessary to reduce qualia? Imagine a mind that could experience all the stuff consciously that the unconscious mind does all the time: Control breathing, heart beats, digestion, coming up with ideas, concocting dreams. If the mind could transfer all these functions to the conscious mind without interfering with performance, why not?
Yes, that’s why “realize” is usually used like. But a “zombie machine” like a thermostat can react in feedback loops to external stimuli. If the machine is complex enough to adapt its behavior to certain circumstances, it would be though to have acted according to certain “realizations”. Qualia may not be necessary for that.
Yes, that is the question. The core question is what kind of complexity gives rise to qualia.
Sensory overload. People with savant syndrom experience only partial sensory overload and many of them can´t live without help. The reduction of qualia helps us to focus on things that might be more important; for example what our conversational partner says instead of every other human being surrounding us within earshot in a crowded environment. If we would actively listen to everyone, our ears could hear, we would not stand one minute in a crowd without a nervous breakdown. But i think that everything is nevertheless part of our memory, even if we don´t have access.
(The function to forget is well known in psychology when it comes to traumatic experiences.)
Yes, why not. With technological brain enhancement this might be possible, someday.
It is just my definition of “unconscious”. You define “unconscious” as zero qualia and i define it as little qualia.
If qualia is a question of complexity. It seems very likely that at a certain level of complexity qualia emerges. But it is also possible that it is a completely new quality without the necessity of complexity…and that some spark…or particle is enough so that even simple organisms like amoeba are conscious if they possess this particle or whatever.