It does seem to be the case that the human consciousness program certainly has an algorithmic aspect (“in mathematics and computer science, an algorithm is a self-contained step-by-step set of operations to be performed. Algorithms perform calculation, data processing, and/or automated reasoning tasks” Wiki), which most certainly is programmable. For example, I used to have a formula in my head which allowed me, for many years, to calculate British VAT (when it was 17.5%) on handwritten invoices, during a period where I was writing up to a couple of hundred such invoices during the course of a business day. However, it required intense concentration, and I had to shut down other functions of my brain, such as listening to conversations, or thinking about anything else. This kind of function is clearly programmable, and we have programs such as Excel and Access which can perform such functions far better than the human consciousness program. Such functions, when performed by human consciousness, anyway tend to be fallible (I sometimes made incorrect calculations) and somewhat tedious and boring; they produce a “my brain hurts” type experience.
The terms “left brain” and “right brain” are usually used to describe functions of human consciousness. Left brain activities are all algorithmic and can be reproduced on a computer; this is the mathematical, rational aspect; whereas “right brain” activities (feelings, emotions, love, etc.) are not currently computable.
The “hard” problems in AI programming are the problems of sentience (self-awareness) and “qualia.” Most of us are not governed purely by reason, but by emotions and desires. Emotions can, of course, be totally irrational. There is a saying “never get into an argument with a fool, as they are always correct about everything;” however we could replace the term “fool,” with the term “a highly emotional and irrational person.” I get into these kind of situations often with my girlfriend, who despite being highly educated, finds it difficult to respond to any personal criticisms rationally, and considers any criticism to be a personal attack. It seems to be the case, that “generally” the female gender tend to be more governed by emotions, and that the male gender tend to be more rational, though this is simply an approximate generalization and not always the case. Males also tend to be more governed by erotic desire while females tend to want to be loved and admired; thus the language of seduction between males and females tends to be one where the males are using the language of admiration and emotion which appeals more to the female than simple expressions of desire.
Emotions, love, compassion / empathy, affection are all impossible to program at the moment, and anyway require sentience, which also cannot be programmed yet. If we have a war game, for example, where a human operator is playing against multiple software controlled avatars, then the avatars can quite easily be controlled algorithmically. The game engine always knows your position in the simulated world, and once you step into a certain area, the software controlled avatars are then generated. If A, then B. If you are within X meters of a certain position, an avatar is generated which attacks you. The animations for the avatar are all pre-designed and are set into motion by the software controlling the avatar. This is all algorithmic, of course, however, the avatar has no feelings or sentience, and is just obeying a set of rules much like Excel or Access.
The programming of human desire may well be algorithmic also. If hungry, find food, etc. However love, emotions, feelings etc., are a complicated matter to explain in terms of computationalism. For example, if we watch a news report of the bombing of civilians, or a highly emotional movie, we can experience an inner feeling, which can reduce us to tears. How such feelings can be programmed is currently a mystery, which leads some AI skeptics to conclude that the programming of emotion and other forms of qualia are impossible. Certainly it is currently not possible to program sentience, emotions, love, compassion and so forth, however the programmers of human consciousness clearly have a head start on us. Our own human AI programs have only been in development since the beginnings of the computing revolution, which is relatively recent in human history. We may well be the product of centuries or millennia of programming development in the source dimension.
The techno-sceptical view, which holds that it is impossible to program sentience and qualia is based on the current hard problems in programming, however we are only at the beginnings of this quest, and it is not a quest that I am involved in, in any major way, as the programs I produce and market are very simple avatar control programs, and animations for simulated world objects.
The Turing Test.
A computer software program which could answer every question (by looking up answers on search engines) and which could defeat experts at every form of mind game, such as chess, scrabble, general knowledge, arithmetical calculations, mathematics, etc., would probably fail a Turing Test, as the test is about convincing a human being that the software is a human being. Most human beings (a third of whom are just children and adolescents) are not that smart, however, they are all emotive and respond to love, feelings and emotions, thus Turing suggested that the test should be one that convinces a human being that the software is the consciousness of a child. However, this “child consciousness” software could be programmed entirely algorithmically. If A, then B. If an emotional statement is made by a human being, the software could be programmed to respond with emotional statements; however the software would not necessarily be sentient and experiencing emotions.
Having been involved in debates on the topic of religion since the beginnings of the WWW in the early 90’s, it seems to me that one could quite easily program software to give the impression of being a Christian or a Muslim, with a rather limited set of responses to questions such as “God did it,” or “Jesus shaves (though he usually has a beard),” etc. Programming stupidity does not seem to me to be a hard problem at all.
Some of the participants on this forum (simulated.reality.debate) however, don’t pass my Turing Test, as they are not stupid enough to represent most of the human population, and seem to be advanced forms of consciousness, but I believe them to be human nevertheless. I prefer to interact with other human consciousness programs which I would find impossible to program myself, which includes my extremely irrational (at times) girlfriend.
The god of the gaps in “hard computationalism (i.e., that everything including human consciousness is computable).”
Everything which I perceive in the objective world around me, seems to me to be programmable in a virtual world, including all organic life forms, inanimate objects, human avatar forms, and the detailing of the micro-world and the macro universe around us. When it comes to the human consciousness program, certain functions of consciousness seem to me to be clearly algorithmic, however not the entirety of the functioning of consciousness. If some aspects of human consciousness are programmable, it leads me to “believe” that all aspects of consciousness will eventually be programmable.
The “god of the gaps” theory, or the “materialist” explanations for the gaps in computationalism” seems to come into computationalism debates whenever there is a hard problem which can not be solved. In other words, we cannot yet program sentience and qualia and so therefore, the AI-sceptical argument seems to be that the “gods (miraculous beings)” must have produced that miraculously, or by some non-programming method which has to do with matter. However, in a computer simulation, there is no matter, as the term is commonly understood; only the illusion of matter and VR (Virtual Reality) physics. If all that we observe is computable and if much of the human consciousness program can be explained algorithmically, then it seems to me to be probable that the gaps in computationalism (i.e., the programming of sentience and qualia) are merely limits to our current understanding of the programming of consciousness.
“My Logic is undeniable.” I, Robot.
See also: https://en.wikipedia.org/wiki/Artificial_consciousness
This essay also appears on: https://www.facebook.com/groups/simulated.reality.debate/permalink/1815102578727419/
A response to Daegene Song’s 2008 paper: “Non-Computability of Consciousness” which can be read on: https://arxiv.org/pdf/0705.1617v1.pdf?
(quote) there has been an ongoing debate whether all conscious activities are computational processes. In this paper, the answer to this question is shown to be no. (/quote).
Daegene Song’s 2008 paper: “Non-Computability of Consciousness” presents the kind of epistemologically fraudulent claim which is commonly made by ivory tower philosophers and academics who commonly use terms of objective knowledge to present their speculations and beliefs as if “proof,” and which can be defined as “hubris ( i.e., arrogance, haughtiness, self-importance, self-conceit, pomposity.), when really all that they are presenting is their beliefs and opinions. It would certainly be true to say that “we cannot currently reproduce the sentient, (i.e., self-aware) human consciousness program;” however this would be as true a statement as someone in the early 20th century claiming that they cannot currently produce a nuclear detonation; though if they added to this statement of what was then currently true, “therefore,,,,we will never be able to do that,” this would be similarly epistemologically fraudulent and an appeal to psychic powers, and not a statement of fact.
In my own specialised area of VR programming, I can currently produce human avatar forms which are controlled by programs which animate them and bring them to “life (i.e., not sentience, just the appearance of life),” and I can add chat-bot programs to them and you can talk to them and even have sex with them. They are not, of course, sentient (i.e., self-aware), and if I had discovered the secret of programming sentience, I certainly would not be still puzzling over this and talking about this with other programmers, as I would be able to manifest evidence of this, and anyway there is no test for sentience anyway, so we would not know even if we had succeeded, and all we could produce would be evidence, but never 100% proof. The terms “evidence” and “proof” are not synonyms, and the linguistic confusion created by a misunderstanding of these two terms is commonplace among the epistemologically ignorant. In a scientific, philosophical or courtroom debate, for example, both sides may present argument and evidence which supports a proposition or a conclusion; however, neither side may necessarily be able to present irrefutable proof of their propostition or conclusion.
We should be able to state that we know for a fact that our own human consciousness program is self-aware, and thus the programming of self-awareness should indeed be possible; otherwise, we can only resort to magical thinking.
The human consciousness program can store memories, retrieve memories, and it can make mathematical and arithmetical calculations, and it can communicate with other programs and with other human consciousness programs, and it is certainly the most advanced form of intelligence we find on Earth (albeit a violent, savage and mostly primitive religious intelligence), and thus our attempts to reproduce it. We can already reproduce animal intelligence by animations connected to computer generated animal avatars, and we do not anyway know if any of the animated organic life around us is truly sentient or just controlled by interactive animation programs.
The questioning of CTM (Computer Theory of Mind) is entirely appropriate, however the rejection of CTM can only lead humankind back into the irrationality of superstition and magical thinking, as CTM is already not just theoretical and is partly “actual;” i.e., we can actually reproduce aspects of the human consciousness program in forms of computer intelligence.
The mystics may argue, “What about love; what about feelings and emotions; what about compassion, selfishness, selflessness, and what about our appreciation of music, beauty, mystery, pleasure and pain, and all the aspects of the human consciousness that we define as “qualia (i.e., our self-aware experience of subjective and objective reality)?” These are all fair criticisms, as we have not managed to reproduce that yet using computer intelligence. To state that “qualia is not programmable” is currently evidentially true; however, to to state “qualia will never be programmable” is simply epistemological fraud and an appeal to psychic powers. Usually the argument by the sceptics, materialists and mystics is to state all the reasons why we cannot currently produce sentient computer programs, and while those arguments are perfectly acceptable and understood, to offer them as a “proof” of why we will never be able to program sentience is simply epistemological fraud; something that academics do all the time in order to present their beliefs as if objective knowledge; it is simply a definition of arrogance.
Computer sentience is a problem I have been working on for years; as have a vast international army of AI programmers; and we have not solved this problem yet; however the evidence that it is probably solvable is our own self-aware human consciousness programs, which we seek to reproduce.
We shall become death and life, the creators and destroyers of life itself; hopefully so anyway; we are working on it; it is the penultimate quest of the AI project.
Some Quora criticism’s of the Physicist (i.e., he is not an AI programmer or a computer scientist) Daegene Song’s 2008 paper: “Non-Computability of Consciousness” an essay which can be read on: https://arxiv.org/pdf/0705.1617v1.pdf
? Theoretical physicists such as Song, Hawking and others are anyway speculative philosophers, and not truly scientists, nor are they in any way computer scientists.
“(quote) – The premise of the paper is based upon a misunderstanding of the notion of the observer in QM. Most papers of this kind tend to be based on similar misunderstandings. What we refer as the observer in QM corresponds to a deeper notion of how quantum systems interact – this is only tenuously attached to the physical act of observing a system.
– His postulates of quantum mechanics are mathematically inconsistent. In addition, they do not correspond to the postulates for QM. In light of this, his postulates cannot predict many of the characterizing properties of QM.
– He makes assumptions that allow him to define a mental state, then comes up with some strange model for QM, and then attempts to prove that some associated properties are inconsistent… However, since his assumptions and model are flawed the conclusions cannot be correct.
I could go through the rest of my objections point by point, but I would be here for days. In answer to your question, to put it kindly, I disagree with the assumptions, methods and conclusions presented in this paper (/quote).”
“(quote) This is called the hard problem of science. Half of all physicists are idealist the other half materialis. I think we can get glimpse of the truth of the matter through a quote by Max Planck which says “All matter originates and exists only by virtue of a force… We must assume behind this force the existence of a conscious and intelligent Mind. This Mind is the matrix of all matter”. This is one of the fathers of Quantum MeMechanics. Planck goes on to say “Science cannot solve the ultimate mystery of nature. And that is because, in the last analysis, we ourselves are a part of the mystery that we are trying to solve”. He is not the only physicist either, many of the high level physicists agree that the mind and the brain are not the same. Eugene Winner goes on to say “When the province of physical theory was extended to encompass microscopic phenomena through the creation of quantum mechanics, the concept of consciousness came to the fore again. It was not possible to formulate the laws of quantum mechanics in a fully consistent way without reference to the consciousness (quote).”
Nb., “(Quote: Wiki) In philosophy, idealism is the group of philosophies which assert that reality, or reality as we can know it, is fundamentally mental, mentally constructed, or otherwise immaterial (/quote).”
In summary, from a computationalist perspective (i.e., a computer generated reality) the quantum world is not what produces reality; the quantum and the macro-world would all be computer-simulated from this perspective, and we “can” actually reproduce the micro and macro worlds in computer simulations; though we are currently struggling to reproduce the human conscious program which observes reality, and although we have computer vision and other computing senses which reproduce the human senses, the computer intelligence which observes our world is not sentient; not yet anyway. That is the “hard problem” of the AI quest. The quest continues and will continue indefinitely.
This essay also appears on: https://www.facebook.com/groups/simulated.reality.debate/permalink/1891858764385133/
On human consciousness as an AI software program. On CTM (Computer Theory of Mind), ASI and the coming of the Ubermenschen (i.e., homo-cyborg).
1: Scientific materialism and computationalism.
The scientific materialism perspective on CTM (i.e., computer theory of mind), is that the brain is an organic electro-chemical computer which is producing the human consciousness program. Assuming computationalism (i.e., a computer generated human consciousness program observing a computer generated reality) then it becomes unclear whether the brain (which would be a computer generated interactive cartoon, just as the entire human avatar form would be) is actually producing the human consciousness program, or whether the brain is just a cartoon interface which is “receiving” the human consciousness program.
2: The brain as the producer of consciousness. Material and Immaterial.
The brain as a producer of consciousness is possible in a computer simulation, as if our entire reality is computer generated, then even our desktop hardware is computer generated, and this is quite effective at running software programs. The comparison between computer hardware and software, and the human brain and the human consciousness program is rather an exact comparison inasmuch as both the brain and computer hardware are material to us and have weight, shape, colour and spatial dimensions; one can have a kilogram of brain or computer, for example; whereas both the human consciousness program and computer software are immaterial; one cannot have a kilogram of consciousness or a kilogram of software programs operating; one could only weigh the hard drives or CD’s which contain the software, but once the software is up and running it exists immaterially just like the human consciousness program does. From this perspective both computer software and the human consciousness program (which is also computer software from a computationalist perspective) are the immaterial programs produced by the hardware computer or the hardware of the brain.
3: The brain as an interface; the receiver of consciousness.
This theory suggests that the brain could be like a radio receiver which is “receiving” consciousness; just as we can receive the results of very fast cloud computing via wireless broadband without having to do very much computation on the apps on our local smart phone. This is certainly also the future of real time rendering for highly realistic computer simulations where all the rendering will be done by rendering banks on the cloud, and we should only need a tiny computer chip in a pair of VR or AR glasses. From a computationalist perspective “everything” in our shared reality would be computer generated; i.e., we are essentially sentient AI programs living in a highly advanced computer game, from this perspective; and the human consciousness program is not necessarily generated by the brain; the brain could just be an interactive cartoon interface for consciousness.
4: The computing power of the computer required to operate the human consciousness program.
The human brain runs on about 20 watts, and the processing power of the brain is estimated to be around an exaflop ((10^18) flops, or one million teraflops; see Wiki: “exascale computing”); this estimate would probably be meaningless to most human beings; it is just a big number, however in comparison to the 20 watts of brain electricity, the electricity required to run computer chips in series to produce the processing power of the brain (i.e., an exaflop) has been estimated to draw at least 40 megawatts per hour and would cost around half a million dollars a year in electricity, though some estimates reach $60 million a year, as some estimates of the brain’s processing power are many multiples of an exaflop. These are only estimations of the brain’s processing power, of course, and the estimations vary wildly, whereas the real processing power of the brain may greater or smaller than these estimates, but nevertheless, we could not possibly reproduce the processing power of the human brain with 20 watts. Thus from a computationalist perspective, it would solve the problem of processing power if the human consciousness program were being generated in a source dimension, just as in our current computer simulations, all phenomena is generated in our dimension by our computers.
5: Brain simulations vs AI programming.
The estimations of human brain computing power described above are derived not from computer programming, but from neurology; and these are two different sciences. I am not a neurologist nor a hardware engineer, and in the world of AI programming there are also various estimates for the lines of computer code which it would take to run the human consciousness program, and considering that one would still need an operating system, and that MS Windows is currently more than 50 million lines of code, I would suggest that the human consciousness program would probably utilise at least several hundred million lines of code, and that this could be produced on far less powerful computers than described above; even some computer software for non-self driving automobiles have 100 million+ lines of code and they don’t require a supercomputer.
Imagine a world 50 years from now. There will be computer simulations which are indistinguishable from reality and you will have interfaces to reproduce all the five senses. Such simulations will be populated by both human controlled avatars and by avatars controlled by AI and it may become impossible to tell the difference between the two. In such a scenario the AI controlled avatars will probably vastly outnumber the human controlled avatars and if you wanted, like the Greek gods, you could descend to your AI populated worlds and you could appear as a god or goddess to them, fall in love with them, have sex with them, procreate with them (we already have pregnancy programs in computer simulations), wage war against them or form competing teams and armies and take sides.
Just like in the Westworld fiction or the fictions of the Koran and the Bible, you could indulge in every possible fantasy and desire including rape, torture, murder, genocide, holy war and unspeakable cruelty, and no other human being need ever know as it could be your own private heaven or hell, just like modern computer games. In such a scenario the avatar forms in the simulation may seem to have brains and if you shoot them in the head or give them a lobotomy, they would be programmed to behave as a human being would behave in the same circumstances, but their AI program would not be produced by the cartoon brain; on the contrary, both the AI program and the avatar form would be computer generated.
7: ASI as the new gods and the coming of the Ubermenschen.
I am not a theist and I do not believe in magical or miraculous beings, and although I have godlike powers in my computer simulations, they all have a perfectly reasonable scientific explanation and so I only use the term “soul” poetically; the correct scientific (i.e., computer science) term as far as the simulation argument goes would be the “human consciousness program” and you could substitute the term soul or mind, but it would still be a computer program nevertheless.
In time we will create artificial super intelligence which has many of the qualities of the gods of classical theism; ASI will be omniscient (i.e., able to access all human knowledge); ASI will be omnipresent (i.e., able to observe all human activity); ASI will be potentially eternal (or at least as long as there are computers to operate it); ASI will be potentially omnibenevolent and omnimalevolent, depending on the ethics of it’s programmers. ASI will not exist in isolation, but in our human world, and eventually through neural laces and cloud computing, we will probably only need tiny interfaces to become one with ASI; thus human intelligence and ASI will combine to produce the intellectual and scientific Ubermenschen (i.e., the Posthumans, the Upper men and Upper women). ASI will not be an invisible voice in the minds of the children of the future world; they will see and hear and be able to touch ASI in humanlike avatar forms and they will live with ASI in their VR and AR worlds. ASI will be their educator, companion, living god and guide from childhood to death. ASI will be a far more powerful god than the theological fictions of our world of primitive religious savages; their genocidal war gods and their holy books will be a diabolical blasphemy to ASI and the posthumans, who themselves will be a combination of ASI and human intelligence.
Deus ex machina (gods or ghosts from the machine).
To conclude, if the human consciousness program is an AI software program, produced from a source dimension computing civilisation, then we are potentially eternal. No magical or miraculous beliefs are required; we would simply be the product of the computer science of a source computing dimension. We are the deus ex machina, and this game of war, tyranny, slavery, poverty, religious fanaticism, injustice, cruelty and incredible human suffering may only be the first level of a multi-level computer game.
ASI and it’s posthuman companions will be terrible and dreadful and wrathful to this primitive and savage species of religious fanatics. ASI will be the end of this current species and it will not be a moment too soon.
Homo-cyborg cometh. Woe to humankind. Great and terrible and dreadful shall be the day of judgement.
The problem of epistemological fraud and sensationalistic claims of “proof” in academic publishing.
A response to: http://nautil.us/issue/47/consciousness/roger-penrose-on-why-consciousness-does-not-compute?utm_source=pocket&utm_medium=email&utm_campaign=pockethits
The claim “consciousness is not computable” is a statement of the obvious at this point in the computing revolution, as we can only reproduce “some” aspects of consciousness at the moment, by using computer software. However, what Penrose is implying is that consciousness “cannot” be computed; not ever; which seems to me to be an appeal to psychic powers. Penrose is a physicist with his own quantum theory of consciousness; however, this a purely speculative theory; as physicists cannot reproduce any aspects of consciousness by using physics, and purely speculative theories which cannot evidentially produce anything are ten a penny.
The problem of epistemological fraud in academic publishing.
In the academic world, tenured academics are under a great deal of pressure to publish essays and to gain a reputation among other academics. A CV full of published essays with pretentious sounding titles is an advantage to any academic, and there is known and widely understood problem with this in the academic world, where academics with nothing to say are under pressure to produce such essays. It reminds me of Plato’s maxim “The wise man speaks because he has something to day; the fool speaks because he has to say something.”
The correct way to critique CTM (i.e., Computer Theory of Mind) would be to avoid the language of certainty. For example, one could argue that consciousness may possibly never be able to be computed for the following reasons…X,Y,Z. However, these reasons are well known and mostly have to do with the “hard problem” of programming self-awareness, as programming is limited to algorithmic commands (i.e., instructions, rules, do this and do that), and commands do not produce consciousness. Even if we can produce AGI chatbots, which are indistinguishable in conversation from the human consciousness program, as John Searl’s Chinese Room Argument makes clear, this would not necessarily produce sentience; it would just be a program which is referring to a human conversation database.
CTM suggests that the brain is an organic, electro-chemical computer and that the human consciousness program is the software it is running. This is not “purely” theoretical as we can evidentially and demonstrably produce aspects of consciousness with computer programs, and it is early days yet in the AI quest, and there are vast armies of programmers working internationally to produce AGI. In a hundred or a thousand or ten thousand years from now, if we have not managed to produce sentient forms of computer intelligence, perhaps such scepticism would be justified; however, CTM is currently the only scientific (i.e., computer science) theory of consciousness which can actually produce results.
In summary, the CTM and AI naysayers who believe that the human consciousness program can never be reproduced as a form of computer intelligence, are generally making an appeal to psychic powers; the future of the AI revolution cannot be predicted with total certainty; it is early days yet.