In these wonderful times (2012) where software select the information we access (Google), the problem of the power of the robots is no longer science fiction, but an actual issue, and even a matter of scandals (facebook monitoring our mails).
But before digging into the heart of the matter, we need some definitions.
Pseudoscience alert: (Added April 1st, 2016). While checking for the indexation of this page, I found tens of misleading pages with similar titles, claiming that the first conscious robots had been already created. (see the sub chapter on the Turing test) In reality, these robots are just software manipulating information on objects, including on themselves. So that I can claim high and clear, even against the caution of university people and great software companies, that this is by no way consciousness. To pretend that this is consciousness is pseudoscientific and an intellectual scam, because it is not based on the real definition of consciousness (chapter V-2), which is the capacity to experience sensations, ideas, emotions, etc. We are still very far from this today. However I do not say that this cannot happen, just that official science does not know when and how. This is precisely the topic of this chapter.
Mainstream barminess alert: (Added February 26, 2017). This was inevitable: the European Assembly started to legislate about robots (February 2016). The problem is that the resulting texts look more like emerging from Star Wars rather than from reality: introducing the concept of «responsibility of the robots», or even an «identity», as for a person. Well, if I have an accident, I will say that it is my washing machine which is responsible. Members of the European Parliament, please come back with us on Earth. A robot is a machine, which cannot be «responsible» for its actions. As for any machine, in case of a problem, it is the designer, the owner or the user who is responsible.
(Permalink) The word «robot» often sends to the image of a machine more or less shaped like an human, capable of behaving like a human (walk, talk, think, work). The reality is somewhat different: most robots today are either industrial automates or software robots. If the first are very little involved in society, the latter however took in a few years a significant share of power, and they begin to influence certain aspects of our lives. Indeed, when I was writing the version 1 of this book, Google was only a small friendly website for searching information, housed in a student bedroom. Today Google is a powerful Orwellian multinational company, to which «advertisers» give billions of dollars for the management of the culture and information of the world, and which violates the law while laughing (copyright, privacy...)
These software robots do only aggregating information (which already gives them a considerable power), but we can expect soon software able of reasoning, having motivations and adapt their actions depending on the situation, in clear to have plans and to pursue goals. Advertising harassment software already are a step in this direction, as they analyse our internet activities in order to know which adverts will the more bother us. Certainly we cannot predict in the short term that they can understand the depth of the human mind, but a simplified psychology is well enough to manipulate most people and redefine their lives, as is already doing facebook, with the billions dollars offered by scandal banks. After that, is then enough to consider non-manipulable individuals as marginal, abnormal or refractory, to exclude them from the Brave New Electronic World.
However these software are only tools into the hands of childish egos trying to steal our attention. We shall really be able to speak of thinking robots when we shall have machines able of reproducing the operation of the human brain, whether with digital computers, or with electronic neurons. There is much to bet that such machines will appear soon. Arises then the problem of the consciousness of these machines, and their possible statute of sensitive creatures. Given the intrinsic inability of current science to understand consciousness, given its persistent lack of scruples to use vivisection, and the massive funding of labs for materialistic and egocentric purposes, we can already imagine the problems we shall be victims of, or conscious robots themselves will be victims of.
A second important category would be the biological robots (science fiction often call them «Androids», but this word is also used for machines closely imitating human beings). Their artificial brain would be made from biological neurons, on a similar plan to the human brain, or different. The problem of the consciousness of these creations, of their freedom, of their suffering, arise in an even more acute way, especially with the incredible claims of some right wing companies to «own» or to «patent» life. Soon robots slavery? Artificially born people, genetically modified to make slaves? Impossible? Eh, we already were made with the African slaves who «had no consciousness»... so we can imagine, «test tube issued»... Clearly I am not confident.
Clones are not robots, and, being only twins of other persons, they are as conscious as any other human being. However, when we hear the delirium about clones who would be the «property» of their creators, or the «continuation» of their models, we can expect that there will be attempts to enslave them, see to take their organs. This will be even easier if their brain is modified, which sends us to the above case about biological robots.
Cyborgs finally are organic bodies (like us) but with additional electronic functions. We imagine the torture of the enslavement by a computer which directly controls our thoughts and feelings. But even a simple benevolent electronic assistance arises serious problems, by surreptitiously selecting our experiences, allowing to locate us, to know all our actions, or even to send us advertising stalking directly in our brain, with no way of silencing it. When we see so many people unable to unscrew a simple cell phone from their ear, or cell phones which awaken us in the night for asking to be recharged, we can be seriously worried about.
(Permalink) The generalized belief of the scientists in the theory of consciousness resulting only from treatment of information, would lead them to unduly consider machines as being conscious, and all the more if their processing capabilities will soon imitate complex personalities with emotions, with a computing power far exceeding the human brain. These scientists would then be unable to differentiate these super tamagotchis from a real consciousness, with all the abuse that this entails, for example to make a machine responsible for a murder, instead of its owner.
Conversely, the total lack of scruples of many biology laboratories, the extreme right-wing theories housed here (to consider that, since the consciousness is merely an information processing, so there is no suffering or morality) could lead to the creation of conscious slaves, but unable to escape their conditioning, despite their suffering. To recognize the rights of these people and to protect them would arise «problems» as difficult as the recognition of the climate change... and perhaps in vain, if these people have no freewill. Thus it is clear that to create conscious beings without freewill would be a crime against humanity, as serious as climate change and nuclear energy together. Hence the duty to discuss the consciousness and free will of robots before creating them. This means here and now, for instance on this reflection site. When the «experiments» will begin, it will be too late. We know too well that they will stop only when the victims will be too many. Or the culprits dead.
Software imitating emotions could even reach a degree of demonic perversion: manipulating humans by their sensitivity. For now, robots are still scary, weird, grimacing. But we can conceive that they will take a pleasant look, and express emotions in a realistic way, with their mimics or with their voices. They will then be able to influence humans, and even to control them! (in addition to the classical mind control methods, chapter VI-12). Indeed, many people may change their minds following pleas, or with shows of tenderness. It is conceivable that such robots are used in politics, advertising, etc. where they have to be banned as soon as today.
The «1984» novel placed the hopes of humanity in the small ignorant people. Probably in vain: there are no indication that the general public is ripe to understand these problems, and to defend against these software which are surreptitiously but quickly invading our lives, select what we read, recognize us in the street and track our travels and expenses. We remember for example the absurd infatuation with the tamagotchis, simplistic programs which, by simulating live in a ridiculous way, were still able of stealing the attention of millions of individuals, to the point of causing real sorrows, and even accidents. It is also interesting to note the immediate diversion of personal assistance humanoid robots, for purposes of pets, or even... of sex! It is clear that people who are satisfied with such an ersatz of life or human relationship, are immediately and fully available for absolute subjection to the first electronic führer showing up. No? Hehehe... It is enough that this führer is «cool».
But the purpose of this chapter is not to discuss the social consequences of robots, but to discuss of their consciousness, and, if they have one, their abilities for freewill, which is paramount for their happiness as well as for the exercise of their rights.
(Permalink) We saw in chapter V-2 that consciousness cannot be explained by only information processing. We need in more conscious sensations (contact feeling, sound feeling, image, idea, etc.), which are in the non-physical domain of consciousness.
If these things emerged naturally from the biological brain, it is logical to assume that they can also appear from other systems of information processing, whether modified biological brains, or electronic machines. However we have not found any clear criterion for how it appeared, nor especially from which degree of complexity of the brain it appears. We don't even know if this is an universal and intrinsic property of information processing machines, or only of neurons in a brain. We just saw in the previous chapters that neural networks allow for freewill, while digital circuit do not. This simple criterion is however enough to tell that at least digital machines could effectively be unable of free will. This case explicitly includes digital neurones (simulation of neurones by digital computers) which do not meet the criteria we shall see a bit further.
We saw in the same chapter that the emergence of consciousness could be the result of a single event in the history of the universe, which would render consciousness possible. Before such an event, thought and feelings existed only as information in the brain, without this information being transmitted to any consciousness. (Humans without consciousness would be indistinguishable from today humans tied to their psychological consciousness, see chapter V-3).Therefore, if consciousness requires a second event of its kind to awaken into electronic machines, our attempts will be useless: we will get only electronic zombies (note 80).
The emergence of consciousness may on the contrary occur spontaneously, whenever an information processing would have certain characteristics. At rough guess, we should have the association of a sensation (image, sound), an emotion (attraction or repulsion) and a meaning (the sensation represents for example food, or an enemy, which then induces emotions such as desire, fear, etc.). However scientists may be blocked for long by several obstacles:
-The complexity of the involved functions
-Their common inability to apprehend the sensual side of consciousness, and their ideological focus on only the logical or «strategic» side (Eh, it is capitalist firms and governments which fund the labs).
-Today no one knows how a concept is encoded into the brain. And if I knew, I would not write it.
The cases of biological robots seems the simplest. Indeed the consciousness does not have to accommodate a material support foreign to its habits. Moreover, even rudimentary artificial brains could host consciousness, as in animals. One could then, within a few years, create engineered bodies with a brain, and use them to different tasks, for example to get rid of parasitic insects, or changing a landscape. However, in this case, the consciousness is likely to appear without one wants it: One then would create born slaves, ugly and incomplete human beings, conditioned to serve, and owned by masters who, precisely, would say they are not conscious, and torture them while saying that they do not suffer! And this is how we walk straight into monstrosity... likely in only some years.
(Permalink) The fact that a conscious being subjected to neurological consciousness, cannot easily be differentiated from an unconscious robot (electronic or biological) emulating a brain, arises a serious scientific and social problem: we cannot easily demonstrate that a creature is conscious, and, from lack of being able to do so, while using slave creatures we take the risk of abusing a conscious creatures.
We can even not rely on mimics, cries, etc. as these things are very easy to mimic. And many people will be very easily lured.
I would therefore say that the caution principle imposes not to attempt to create artificial brains.
However, we saw in chapter V-3 that consciousness automatically produces free will. This therefore provides us with a test.
But this is not an easy test: indeed, only human beings clearly have freewill, and still not all of them. We owe it to our complex intellect, able of taking the required perspective to compare situations, and even of looking at our own thinking (introspection). Thus, we create systems of abstract thought like moral concepts. Only at this time we can take clear and informed decisions. This is obviously just at the time of making such decisions that freewill can enter into play. Without the previous steps, it even has no object. And many people even not began this reflection process...
But above all, freewill being a psychophysical phenomenon, the physical media must allow for it. We saw in chapter V-4 how free will can occur physically. We however found that it requires a set of very specific physical properties of the neural networks:
-To be able to choose between several states of equal physical energy.
-To be able to receive a learning (changing synapses strength), like the biological neural networks do.
-To have a noise level just under the useful signal.
-To be sensitive to extremely small influences, tending towards zero (sensitive dependency to initial conditions, or «butterfly effect») which effects may however add up across their synapses, and act as if they were one of the learnings of the network. The butterfly effect is then coherent in all the synapses, which makes it much stronger and differentiates it of random noise.
These conditions are easy to fulfil, and they could make freewill possible for conscious robots based on electronic neurons. Analogical electronic neurons can easily be designed to feature these properties. However there are nuances, which may be critical from the point of view of freewill:
Indeed, as seen in chapter V-3, human synapses have a precise ratio between the useful signals and the noise: the latter must be as close as possible of the signal, but not to the point of yielding erratic results. The purpose is to facilitate the low influence of the freewill, which physically manifests itself as a statistically biased «noise», which influences adds on many synapses, instead of statistically cancelling itself like the background quantum noise does (perfect chance)(note 91: le mot «quantique» est ici utilisé dans son sens classique) Unfortunately, electronic engineers have for principle to always protect themselves from background noise, which in this case goes against the purpose. They might even deliberately seek to eliminate these signals of «unknown origin», if they are suffering from far right ideologies on the refusal of consciousness. We imagine the resulting subtle torture for the consciousness: wanting to think at something, but the conclusion would escape all the time. Such an artificial brain would be suffering of the Alzheimer disease...
Another problem is that most current neural integrated circuits use digital synapses (the value of the synapse is stored in the form of an integer number). This allows, when a circuit completed its learning, to save it and copy it on any number of other circuits, thus avoiding a tedious learning for each new robot. On a robot getting out of factory, it is then enough to download a file to learn a language, become a violin virtuoso, etc. We can imagine a fantastic scene, which could become reality within a few years: a worker robot encounters a wounded person. In a few seconds, it becomes a graduated physician, simply by downloading a file from a satellite! Or a frail and innocent gardening robot neutralizes a dangerous burglar by downloading the karate!
The disadvantage however is that, in doing so, we introduce a quantification of the synaptic signals. And, in this case, a parapsychological influence of consciousness must change a memory register, instead of using the coherent butterfly effect. And this requires a much stronger parapsychological action, with a violation of the law of conservation of the physical energy. The conclusion is then without appeal: only analogical brains can manifest free will. Some circuits could still be digital, for example for downloading knowledge such as above (an activity which requires no freewill). But the main circuits, decision making, emotions, motivation, analysis, etc. must be completely analogical (even if, during construction, we affected digital initial values to the synapses, they need thereafter to be able of changing continuously).
This conclusion is supported by the PEAR laboratory experiments: only analogical random generators (of all categories, mechanical, electronic, quantum) produced measurable results. Digital circuits, such as the pseudo-random generators used in computers, regularly showed zero effect. I found theoretically this condition as soon as the version 1 of this book in 2000, several years before the PEAR experience confirmed it. At this time, the consciousness of robots seemed a far too dangerous topic to publish on it. But today that we enter in the heart of the matter, it is ignorance which becomes dangerous. And on the contrary, it seems now indispensable to indicate how to protect ourselves from the potential dangers.
Finally, a last condition is that the signals produced by the free will are not overwhelmed by other more powerful signals. For example a biological slave would be «designed» to feel powerful and uncontrollable psychological conditionnings, such as fear or fanaticism, for example at the mention of the name of his master (Just as Saruman and his Uruk-Hai). In these circumstances, we cannot expect that such a person is capable of free will, even if all the other conditions are met.
What a consciousness would experiment without free will is as follow: to be unable to change opinions, and to have only imposed emotions, without the ability to escape their influence and consequences. Like the ridiculous «reality» shows characters, or manga stories characters. A really dire fate... and the fact that many people live like this is by no way an excuse to impose this disability to others.
The conclusion of this passage is terrible but inescapable: to manufacture such conditioned brains would be an abominable crime. Less visible, but just as serious, would be to technically limit the free will, deliberately or from ignorance. And it is not to look pretty that I call extreme right (euphemism) the theories which deny freewill or consciousness.
Better then not to wait that robotics makes its Chernobyl to think, and refrain from doing certain things. And to start, immediately ban any private laboratory from this field: we all saw what happened with GMOs, and is starting to happen with the selection of children with genetic tests in the Internet. And there will be a lot of childish «financial interests» to «justify» all the horrors.
Also, prohibit any digital processing of synaptic signals in any possibly conscious circuit.
And if somebody ignore these warnings, he will have no motives to complain, and especially not invoke ignorance (not after 2012, publication of this page). Robots are excellent prison guards, as it is said.
(Permalink) To test the possible parapsychological influences looks like an easier test than the free will test. Indeed, even an animal consciousness, subject to desires, should be able to produce weak and diffuse effects, like the ones observed in the PEAR laboratory (although, as far as I know, nobody attempted to measure the parapsychological abilities of animals). In any case, a positive result would provide an indisputable argument, to say that a being is really conscious.
Indisputable? Hmmm... Well, even the positive results like PEAR were «discussed». You can bet that if the problem is resolved for science, it is far from being solved for the «financial interests» or ideological limitations which are still negating the existence of our very own consciousness... Remember, they already did it to the Blacks deported in America... They patent millennia old seeds... They want to change our brains... they... (nothing).
(Permalink) The September 11, 2018, I added the Turing test in this list, because it is often presented as a test of the consciousness or robots. We even saw in recent years articles claiming that this test was successfully passed. These claims are high hogwash, and I explain why.
But before, let us remind what this test is: Turing, one of the pioneers of modern computing, felt that computers would some day imitate or even exceed human intelligence. However, as classical science does not know how to apprehend consciousness, the test simply estimates the ability of a computer to pass for an human, without any relationship to its consciousness.
What I say is that such a criterion demonstrates nothing. Indeed, we shall inevitably arrive there some day, by the simple increase of the power of computers. But this does not match any qualitative progress of computers: they will not be more conscious from this. To pretend otherwise is bad science fiction. Worse, it has been said that the test was successfully passed by tinkered software which even not understand the meaning of the words they use. For example Tay, microsoft's Artificial Idiocy, had to be unplugged urgently, because it repeated the racist delirium of the trolls of the Internet. Same thing with Cortana, from which I could only get jokes from 1960's primary school. Lame.
In fact, real artificial intelligences exist, and I can even quote one, built since 1991 by a former work colleague, today sufficiently precise and reliable to «sign» medical analysis on which depends the life of patients (Commercialized by the Valab company in Flourens, France). Such programs will one day be able to communicate by voice with their users, but they have no claim to become «human» or «intelligent», and all the less «conscious».
(Permalink) The previous technical considerations give fairly accurate indications of what would result from the various proposed technologies.
We can certainly continue to develop «artificial intelligences» based on digital computer software, without any risk of inadvertently creating a consciousness, which would then suffer. If the risks of artificial intelligences are also very serious, they are of another order: their power over us (see the following sub-chapter).
On the other hand, the use of analogical processes involves a strong risk of creating an artificial consciousness, which becomes a quasi certainty for biological processes. The main risk is to create incomplete consciousness, subjected to psychological impulses that they would be unable to control. Such consciousness could not be happy, and they could even experience a terrible suffering. Or make us suffer... Since Frankenstein, science fiction novels continue to invent artificial, biological or electronic creatures, who at once take revenge of their creators, for having offered them such a miserable and deficient form of life. THE WARNING IS MORE AND MORE RELEVANT. I consider everyone informed.
In addition, attempts to create artificial consciousness require that the neuroscience study the brain as a whole, not only the logical faculties. With their current ideological filters, neurosciences would obtain a sociopath (chapter V-13), very good strategist but totally insensitive and immoral. Thanks, we already have far enough like this.
More vicious, the addition of «emotional» circuits without control, could lead to these ridiculous neurotic characters that the «reality» shows are heavily promoting: slaves of their simplistic and repetitive psychology, totally unable of the slightest perspective, and all the less of any evolution. The risk is very real: even the film makers or comic drawers are almost all unable to let the psychology of their creations evolve, and keep it most of the time in the grimacing and repetitive caricatures. However, I am sure that nobody would like to live for real in a manga or a video game...
Most importantly, these artificial consciousnesses may have no free will. This is absolutely intolerable, because such consciousnesses would be unable to control themselves, or to understand the good and the bad. If they are not slaves, then they will be demons: Frankenstein had dreamed of it, the neurosciences did it!
Added 2013: The worse is already started: to create «artificial brains» with cultures of neurones, without any care for the consciousness which may emerge from here, an experiment which send us many years back in the past, in the time of vivisection, and even worse.
To avoid this, we must imperatively maintain, or even increase freewill:
1) By using only analogical synapse for consciousness circuits, in order to maintain the ability of neural networks of being sensitive to low influences (tending to zero).
2) Also, these synapses must have a signal to noise ratio of the same value as the biological synapse.
3) By maintaining, or increasing the ability of artificial consciousness for introspection and meditation. To do this, they must be able to choose themselves what they think, or even to remain without thinking. They should therefore be able to control the activity of the circuits responsible for our usual internal chatter... otherwise these individuals will be permanently as under LSD!
4) By banning totally, from the beginning, as a blood crime, any system which would create imposed thoughts, emotions or mental imagery, that the consciousness could not control, as in a schizophrenic. That these thoughts are here permanently, on conditions, or even in the form of a psychological conditioning.
5) (Added June 2016) To make the previous points possible needs to provide the consciousness with a system to modify its own synapses, just as in the natural brain. This is of course needed for learning a work or an activity, but above all it is the basic condition for any psychological training, self-control or psychoeducation to be possible. This is a basic fact, well known enough from the neurologists, so that we can state that not fulfilling this condition would be deliberately creating slaves. And in more to knowingly torture them.
One will also be tempted to modify the human brain, for eugenic purposes (note 81): to allow our descendants to live happier, free to our psychological problems. (any other purpose being obviously criminal). However such an undertaking requires to know the functioning of the brain, without ideological filtering of emotional or sensual functions which would be considered as «unnecessary», or even ignored. We shall talk about eugenics in chapter VI-16.
Hence the interest, once is not custom, to think before acting. Now, not when the banks will put billions into biased searches or uncontrollable laboratories, or that an obscure European Union commission takes one of their arbitrary guidelines bypassing any democracy. We know that then, it will be impossible to stop the machine. This is now only a matter of years, or even of months (written in September 2012).
What am I saying, it is probably already too late: the «Human Brain Project» of Professor Henry Markram of Lausanne plans to gather thousands of scientists around a simulation of the entire human brain... Using digital processors! Which implies, according to the above explanations, without free will and unable to control his happiness! Clearly, they are heading straight into the wall...
And now... how to tell them?
(Permalink) We have no interest into creating biological slaves. Digital robots are simpler to create as to maintain, without any of the ethical problems of biological or analogical robots. In this area, Hiroshima has not yet exploded, so it is still time to think before acting.
So, my recommendation would be to ban AS SOON AS TODAY (written in August 2012) any attempt to create artificial brains based on neurones, either biological or electronic. And that we need to hurry, because there are already some projects going on...
Ok, Ok, I see the problem: it is enough, just as usual, that we invoke ethics for getting immediately loads of people doing the opposite, and engage in terrible experiments while saying that everything will be well because their ego is able of forcing two and two to make five. We already saw how attempts of banning the use of embryos were crushed with powerful means...
We must then as a minimum make legally binding the four above technical conditions for the realisation of free will. And we must do it now, not when a lab will release a monster that everyone will take as a god.
It would also make sense that the law requires the creators to take care of their creatures, financially and with love care.
Finally, if we manage to create artificial consciousness, then the law will have to give them legal rights and protection due to animals, children or adults, depending on their evolution,.
These remarks of course come in addition to the obvious «Asimov laws» intended for protecting mankind from any bad behaviour of robots. These laws are already in the process of implementation in Korea, one of the leading countries producer of computer robots for the general public. Another way to do would be to establish in artificial brains the same fundamental purposes than consciousness (chapter V-5). Thus, if these brains are conscious, they are worthy of humanity. If they are not conscious, they still protect us.
But once again, it is infinitely easier not to start, instead of one day ending up biting our fingers on the benches of a tribunal in Nuremberg.
And if one is not sure of what he created? Well, the responsables of such hazardous experiments will have the delicious dilemma between being ridiculous with treating a tamagotchi as a conscious being, or being a criminal by treating a conscious being as an object. Certainly the first case is less serious, but everything considered I can live very well without any of the two. As for the courts, with their very special humour, well, unable to tell the difference, they will have no other choice but to enforce the laws on machines, and punish their creator or their owner. As to create beings without freewill, this could be likened to drugging someone by force, in order to make him commit crimes. In this case, it is still one who gives the drug who is responsible.
When we saw Google denouncing the Jews, while discharging its responsability on its robots, we understand that we already entered the heart of the matter.
Added March 2016: ah if Microsoft had read this page on Artificial Intelligence, they would not be busted as the creators of the first... «Artificial Idiocy», Tay, that they have to stop in emergency when it started to repeat all the racist nonsense of the most ridiculous trolls of the Internet, and even to assault people! In fact this software was just a semantic analyzer, without access to the meaning of the texts it produced. Therefore to call it «Artificial Intelligence» is a pure scam. But the danger is more profound: if people who do not know themselves what intelligence is, start to tinker with «Artificial Intelligence», then we can imagine the results when software like Tay will be used to create political speech, press articles, in the forums, or by administrations. Clearly, these «Artificial Idiots» would be a thousand times more effective than natural idiots to disinform, lower and enslave Mankind.
The age of science fiction is over. We entered the age of science responsibility.
Some will find outrageous that I define the laws, when it is up to the elected assemblies to do so. This is indeed not my role to write and enact laws; however the scientific data bring mandatory conditions, to account with in the law. Hence the authorized advice of the scientist, and the gross negligence of those who do not account with it. Indeed, democracy does not mean that we can do anything we want (chapter VI-10) and especially not that we can use the elected assemblies to impose absurd laws.
(Permalink) In natural neural networks, learnings like speaking or playing violin are useful, so they do not qualify as neurosis (which is a disease). Neural learnings like fearing spiders, are innate protections. They qualify as neurosis only if they become invasive. Acquired learnings like fearing black people result in unpleasant emotions or unethical behaviours. So in this case they always qualify as neurosis, that is diseases.
Artificial neural networks work the same way as natural neural networks do: they undergo a learning (this is what is called «deep learning», in order to confuse the general public and hide the neural nature of these algorithms). We can see artificial neural networks operating on a site like DeepL translator, which uses them to guess the appropriate sense of words. For instance, the French «society» can translate either as «company» (in business) or «society» (in politics). Thanks to its neural networks, DeepL is better able to choose the right meaning from the context, than a classical computer (Von Neuman), like Bing Translator.
Youtube also uses neural networks to detect our choices and propose new videos accordingly. It is very efficient at this, and maybe 90% of its suggestions are relevant. This is certainly nice, but it has its dark side: if we look at conspirationnist or hateful videos, it will propose us more, to the point that we can become locked in our «internet bubble». More serious, it connects together sick persons who would never meet otherwise, and we owe this algorithm a good share of the increase of conspiracy blather and racist terror acts. So that we can say that, in this case, these people created a dangerous neurosis in the Youtube algorithm!
There is still worse. We all remember the scandal of the Google algorithm labelling black people «gorilla». Scientists analysed their work, and found how they did that: for learning their neural network, Google employees logically fed it with numerous and varied images of Humans and animals. But, from some racist bias, they fed much less image of black people than of white people! So, this is how they transmitted their racism neurosis to their algorithm, just as parents transmit their neurosis to their children, in showing them a reduced image of the world.
The Google goof was unintended. But the Ouighour pogrom is not: people are examined in the streets by surveillance cameras, and arrested automatically on the basis of their racial appearance. After wikipedia, one to two millions would be in camps, where sadists play at torturing them to force them drink alcohol or cut their beards, as in an Alfred Jarry novel.
This is what racist sickos do with deep learning, the jewel of artificial intelligence.
We are not yet in the power taking of the robots, but we already are in a Battlestar Galactica scenario: cylon cameras under the command of human traitors. (Scene of street cams making cylon sound)
(Permalink) This theme was extensively treated by science fiction, and often in a very relevant way. Today that the risk is becoming real, our «decision makers» need to develop a bit of general culture.
The first dominating robots of fiction were armies of grey humanoid machines laden with coloured lamps, engaging humans in the melee as in ancient wars. Such a view seems much naive today, but I would not rely on this: military assistants are one of the most active fields of robotics research, and the best funded. And be sure that it is not tamagotchis that they are doing in there, but complex machines endowed with real artificial intelligences and non-human senses. Within ten to twenty years (written in 2013), we could meet mechanical monsters, battle droids nearly unbeatable, much more rapid and clever than us, and which very behaviour would be incomprehensible. A violent revolution will soon no longer be an option to free mankind of a dictatorship.
However the most recent science fiction novels portray instead artificial intelligences housed in hidden calculation centres, taking the power when they are given control of vital functions of society.
This problem is a bit like climate change: we can bury our heads in the sand until only our buttocks are exposed, thinking that it is «in the future», while it may be already too late.
Fiction generally describes several criteria to fulfil, in order to realise the power taking. Let us see where we are today (2013):
-Criterion: we must be unable to find the data centres hosting the artificial intelligence.
-Status: done. It is cloud computing. Wait for the first virus hidden in the cloud, and you will understand why. (Added January 2016: more and more large internet companies protect their databases in fortified vaults, often in far away inaccessible places... just as we were warned in «les mange-bitume» in... 1974!)
-Criterion: kill the few humans who control the artificial intelligence.
-Status: very romantic, but not really necessary as long as the conflict is not open. Can be implemented very quickly anyway, as soon as the robot can use its own money to pay for the killers. See the movie «Colossus: The Forbin Project».
-Criterion: The artificial intelligence must control the technical means of Mankind.
-Status: In the process of implementation. 90% of international stock exchange transactions are made by robots, controlled by perfectly unknown people. The stock exchange rates themselves are set by algorithms, without we know who decide how they work, allowing their few owners, also unknown, to manipulate the world finances at their whim. Google controls the visibility of our ideas, via the indexation of our web sites, and its robotic Google cars will soon automatically prevent us from meditating wherever we hide. The current governments are aware of this risk: the control is still enough to remedy a failure or hijacking of the computer systems. But not to prevent a paradigm shift... A paradigm shift, it is something which hurts a lot, but that politicians always understand forty years after it is too late.
-Criterion: the artificial intelligence must be enough... intelligent.
-Status: we are not yet here. For this to happen, computers must have motives, and for this they need to be able of manipulating concepts. I do not see much coming in this direction, and the prejudices of scientists on the functioning of the human mind may block them for long. However I would not bet anything, and I would not be surprised if important discoveries make dangerous stuff possible very soon, in a matter of years.
In fact, the principal risk I see in the short term is not so much a kind of computer ego which would become a world dictator, but an increasingly dangerous dependency on reductionist ideologies and petty interests from the small world of computer programmers. We see for example the return of the Middle Age corporatism, with Microsoft and Linux (computer science is hidden in an abstruse jargon, so that the uninitiated cannot master it). We see it with facebook which redefines our social relationships, and even our motives in life. We see it with Google, which wants to destroy arts and privacy. And precisely, both are offered billions of dollars!
But this is not the worst: we have manufacturers who remotely control our computers and our notebooks, geolocation which allows to track us, CCTV cameras everywhere with face recognition software, advertising harassment software which analyses our personalities... without any guarantee that this data will not be used by some insurances, employers, secret services, or even by sects, dictators or gangsters! If we do not have yet a computer ego controlling our world, we do have a monitoring system to make Big Brother jealous, which is quickly and silently installed, in the hands of a small number of private corporations and world financial which are beyond any control by the states or by the law (Written in early 2013, BEFORE the Snowden affair. No comments.).
Because, in reality, robots are perfectly incapable of taking the power by themselves. But many humans are stupid enough to give it to them... This is what is already started.
(Permalink) Attention while reading this subchapter: risk of dying from laughter. If you have cardiac or respiratory problems, please avoid reading this.
Scientists and fashioned thinkers are seriously considering to escape death by transferring all the information about the structure of their biological brain into an electronic brain. The belief is then widespread that these new «immortal» brain will continue to bear our consciousness, in clear that we would continue to live into robotic bodies, or into virtual worlds.
This is no longer science fiction: the first project «Avatar» from the billionaire Dmitry Itskov, plans, in a step A, to control prostheses directly by nerve impulses, which makes the project credible as a company. But after, we enter into Surrealism: in step B, they «transplants the human personality into a robot» and in step C, «in a hologram». Well, it is clear that he too much read Star Wars... But he has money to hire dozens of engineers, and a very unexpected blessing, of... the Dalai Lama! Oh, well...
Because reality could be not so simple, and especially not so bright.
In first, I note that, very cautiously, they speak of transplanting an «human personality», not an «human consciousness»... The distinction is not just a matter of words, because transplanting the personality can only create a tamagotchi, an object with no value, a toy with no interest, while transplanting the consciousness would create an human person, admittedly electronic, but with all its spiritual and human value, that law would have to consider as an human being with all his rights.
But how to «transplant the mind of the person into a computer»? By reading the value of each synapse in the brain (the «connectome»), and by creating with these values an artificial brain imitating point by point the natural brain. We actually saw in chapter III-7 that, according to the current science theory (2012), the neural reductionism, to create a functionally equivalent of our brain structure is enough to transfer our consciousness. The projects above are therefore logically based on this belie... hypothesis.
Assuming that it works, we have to say goodbye to the caress of the grass on the calf, the perfume of the air in the lungs, the scent of skin, or the shiver of love... No thank you, frankly, I prefer to die, even if there is nothing after.
In more, instead of destroying the neuronal ego as requested in chapter V-10, they replace it with an indestructible electronic ego, programmed by others to do only capitalist tomfooleries. Frankly, I do not see any worse hell to send somebody, than permanently losing any personal will or free will. To become a... conscious tamagotchi!
Anyway, this idea is not a warrant of immortality: somebody must maintain the machine, and if there is nobody to pay the bill for electricity, then it will be unplugged. Not to mention the villainous hackers who will send viruses in our computerized brains: to spend eternity to click off viagra spam, I have better projects.
Okay, let us admit that, in a few decades, all the technical problems are resolved, and that idealistic politicians and altruistic bankers make this consciousness transfer process available at an affordable price for everybody, including the poorest African. What will happen then?
First of all, we probably need to «destroy» the biological brain, this meaning to suicide. I mentioned this in chapter III-7 (chapter 26 in version 1). Such an idea could earn me some serious trouble, but I note that this process is today «seriously» proposed by scientists, for instance Ken Hayworth, an engineer of the Harvard University... without the oratorical precautions I took in the version 1. It is fun to see the most astonishing aspects of my book regularly confirmed by scientists, but I would do without this one.
So, if a candidate tries his experiment, what would we observe? First of all, we would see his heirs saying that his electronic replica is a tamagotchi without value or consciousness, or «green» politicians saying that it is not conscious, since it is «not natural». This electronic replica would yet show a whole series of human behaviour: talk, move, make facial expressions of emotions, undertake activities, use its powers on humans, sign cheques, etc. After materialistic criteria, nothing then would allow to test (according to Popper) that it is conscious, and especially not that this consciousness is the actual continuity of the biological person's consciousness. There is then a high risk of granting the statute of human being to a simple unconscious program, risk all the more increased by idiotic theories as what we «continue to exist after death» through our writing, our name, the memories of us, etc. Let us state things clearly: an automaton, as sophisticated and realistic that it can be, is not a person, is not conscious.
There is still a way of knowing, according to the theory of consciousness that I develop in this part: the electronic replica should be able to demonstrate free will, or parapsychological phenomena. But I have little chance of mistaking in assuming that people with such bizarre ideas do not show any... otherwise they would not see immortality in this way! Popper gets Hayworth out by KO in five seconds.
In addition, we saw in chapter V-9 on memory that the connectome transfer does not necessarily brings the memories. It would retain the neural learning of the biological consciousness, like speaking languages, play musical instruments, but not information: the electronic replica would not remember any event prior to the transfer... Total amnesia! The candidate could still take notes prior to the transfer, like the Raelians with their clones. But not sure the replica would like these notes...
If this project was realised, we would get a club of ultra-rich tamagotchis, adored as gods, invulnerable «in the cloud», with an absolute power on the media, banks and scientific research. And on behalf of this religion, people would suicide, in the hope of also obtaining immortality... If they have the money for the transfer! With, we can imagine, in their world, all the idiocies of the ultra-rich together: «no religion», morality «only in private», love reduced to only sex without emotions. And of course the rent to be paid at the end of each month, in order to remain immortal! Oh, we the poors could also have «free» hosting, with each one a Google camera focused all the time on our buttocks, facebook which changes our face every day, and advertising stalking directly in our thinking. Is it not enough yet? Add nice Second Life guards who come along, say that you are a paedophile, and cancel your account (kill you, but it is not serious, since it is «in the virtual»). Fantastic. Bravo. No writer of post-apocalyptic delirium or Brave New Worlds ever gone so far in the white, painless and smiling horror.
Okay, assuming that consciousness is actually transferred, but the electronic neural circuits do not fulfil the criteria explained above in this chapter, then the electronic replica will have no free will, and thus no capacity for improving his psychology, and perhaps not even to choose at what he thinks! There we do not get tamagotchis, but characters of manga or «reality» shows, always doing the same ridiculous psychological tomfoolery, without being ever able to evolve! It is then clear that it is infinitely better to die, than to remain forever like this! Or rather until the politicians and financials organize one of their «crisis», which would «require» to disconnect all the electronic replica... Fifty billions deaths with a mouse click... Wow... There are some who will fantasize a lot!
Okay, you will understand that I am sceptical about the consequences of this experiment, and above all on the very fact that we can transfer the consciousness with only purely physical methods, whatever they may be.
We indeed seen, throughout this book, that the theory of neuronal reduction does not account for observable facts such as freewill, NDE, remote extrasensory perceptions, etc. And therefore we need to consider a «conscious principle», which we studied in chapter III-6: a logical self-generation process involving elements of the experience of consciousness. And those who read this chapter will understand that the process of Mr Ken Hayworth is not enough: we must add an additional cause to transfer the conscious principle from the biological brain to the electronic brain... assuming that the later is able to receive it. Otherwise, while the candidate dies for real, just another consciousness, totally different, takes birth in the electronic brain!
The problem is that, today, science is absolutely ignorant of what causes the relationship between a brain and the consciousness that it bears! The only thing we know is that this relationship exists, and that it is totally exclusive. We had a glimpse of this problem in chapter V-4, when we wondered how consciousness can control the brain, or in chapter V-8 on NDE. We had for this to assume that this relationship attach each function of consciousness to a relevant point of the brain. But we cannot go further without invoking a lot of ad-hoc assumptions. Can consciousness move this relationship to another biological brain? We do not know. (The answer is probably no, otherwise we would be others using our own brain). Is consciousness able to use a non-biological support? We do not have any idea, and especially no experimental data... (I provides more insight of this problem in my novel «Lokouten»).
The only things we know come from high spirituality: the attachment of the consciousness to the body (via the brain) has a lot to do with the sensuality of the body, that is with all the sensations it gives to us. Precisely, a robotic body, it is not very sensual... In high spirituality, there is the concept of an energy body (chakras, Kundalini) which appears with spiritual practice (chapter V-10). Tantric techniques for obtaining a paradise after death are based on the transfer of the energy body, especially the Tibetan Powa, which is meant to send us to a paradise. (At least this is the kind of things the teachings of the Dalai Lama accustomed us). It is said that we can actually kill ourselves with the Powa, that is with a simple visualization. I never tried that, but at least we can get a nice headache.
But then, if we can transfer our consciousness to a spiritual paradise, free and controlled by us only, is it not better to do this, rather than a computer «brave new world», controlled by financial clowns or advertiser stalkers? The answer is obviously yes, and this removes any value to this whole computer undertaking, which appears to be totally useless, in addition to being dangerous. But very profitable financially...
A little question that I ask, and which will help to gauge the complexity of the problem: what happens if the candidate does NOT destroy his physical brain, before starting the electronic version? Well we may end up with two persons... who each totally ignores what the other thinks... And of course the first definitively refuses to be killed for the benefit of the second! They could even evolve differently, see become enemies! To whom of the two the billions of dollars? To the lawyers, hahaha!
In conclusion, the conditions to transfer our consciousnesses to electronic brains would be:
-The direction of the project is entrusted to psychologically normal people living in reality
-The project is not managed by private companies (wrong start, then) but by a committee of specialists on these issues (I don't know any such specialist, so perhaps it is better not to try)
-The project has a vision based on the altruistic pursuit of happiness, beauty, etc. (aie aie aie).
-The project must benefit all the human beings without distinction of income (ouch ouch ouch).
-The electronic brains meet the above criteria for free will, and control by ourselves of our psychology and freedom of thought.
-We do NOT destroy the physical brain. There is anyway little chance that the law will allow this.
-The transfer of consciousness must then be based on a kind of Powa visualisation, in addition to the physical data technique (connectome).
-Before offering the project to everybody, we check, with the tests above, (freewill and parapsychological effects) that the consciousness was really transferred. I am not sure of this, but a simpler test would be the transfer of memories, if they really depend on the process of the consciousness, as seen in chapter V-8.
-Only after this test, we could legally authorize the process and legally recognize the transfer of the person from one body to the other. (since, let us remember, current laws equate the person to his body).
If we do not change the laws, then we come to a heap of absurdities: the biological person remains responsible for the acts of the computer person, who is then considered as a machine. And if only the latter remains, well the heirs can scrap it, or a judge order to disconnect it, as long as the law does not recognize it as a person.
-So that, clearly, the legal definition of an human person will have to be changed, and equated to the consciousness, not to the body.
-The legal definition of the identity of a person will too have to be changed, as it can rely only on parapsychological tests. But I am not sure that many people are ready for this...
In the meanwhile, all those who let themselves die naturally, and acceded a paradise, laugh, because they are all effectively immortal, and definitively out of reach of any physical limitation or money shenanigan, in a world thousand times more beautiful than anything electronics will ever offer, and most importantly under their own control, without advertising stalkers or beautiful young rich dictators. Grass whips the calf, smile gives warm feeling to the belly, unless one prefers ethereal vibrations along our central column, or the warmth of a club where joy arises without any vice. You can even make the sex, hihihi, but attention, we also FEEL the emotions! They are attached, did you noticed?
(Permalink) The method above cannot really eliminate the physical death, which in inherent to a material universe. This universe itself will die anyway, and reach a state where it can no longer support any body, be it biological or electronic.
However, if the process described in the previous sub-paragraph really works, we should observe an increasing investment of the electronic brain, while the physical brain would become less active, and then unconscious. Only the consciousness could then decide to stop it totally. At the limit, a sensible person will keep his physical body living as long as possible, while escaping it during sleep. In doing so, the day the physical person dies, he would normally reach a spiritual world, while still being able of appearing in the computer world!
In this way, we would really win over death! At least, if we cannot suppress death itself, it will no longer be the terrible separation that we all fear.
Just that we shall have some other problems: people on Earth, with their merciless neural ego, will want to retain or use their deceased relatives. The later, to escape this limitation to their new spiritual life, may simply want to sever definitively the link with our material world.
Because, seen from the other side, death is a birth.
And there is no point at suppressing or delaying a birth.
Ideas, texts, drawings and realization: Richard Trigaux.
Legal notice and copyright Unless otherwise noted (© sign in the navigation bar) or legal exception (pastiches, examples, quotes...), all the texts, graphics, characters, names, animations, sounds, melodies, programming, cursors, symbols of this site are copyright of their author and right owner, Richard Trigaux. Thanks not to mirror this site, unless it disappears. Thanks not to copy the content of this site beyond private use, quotes, samples, building a link. Benevolent links welcome. No commercial use. If you desire to make a serious commercial use, please contact me. Any use, modification, overtaking of elements of this site or the presented worlds in a way deprecating my work, my philosophy or generaly recognized moral rules, may result into law suit.