12. AI (Artificial Intelligence)
This is the twelfth article in the series From Particles to Angels. If you are interested in this article you should read the previous articles in the series in order, beginning with the first (On Happiness).
When the farmer grows his crops, all she does is plant the seeds in the field, and largely allows nature to do the rest. Anyone can make a tree by planting a seed. So that when we ask whether human beings are clever enough to create life and intelligent beings, we should be mindful of what assistance nature can provide. Notice that nature is not clever. It has no plan or intention. Yet nature was able to create intelligent life quite by accident, blindly and yet almost inevitably. Biological evolution leads to progressively more complex organisms, able to do more elaborate behaviours and more sophisticated information processing. Then at some point in the process of biological evolution consciousness seems to just appear spontaneously. In the two Unity vs Robot Zombies articles we arrived at the conclusion that consciousness seems to be somehow automatically associated with certain forms of organisation of matter, as if there is already something in matter predisposed to give rise to consciousness under certain conditions. So all nature had to do was create a suitable body, and the consciousness would simply appear and take up residence once an appropriate vehicle existed. As inseparable as space and form.
"life - - finds a way." ("Jurassic Park" (1993))
You may have seen a plant clinging to a crack in the side of a concrete building. The crack collects dust from the wind, and eventually a seed which then takes root. Nature exploits any opportunity to grow life. We only need to provide an opportunity, intentionally or not. In short, we do not need to know how to create consciousness in order to create consciousness. All we need to do is create a body that can serve as an appropriate vehicle for a consciousness, and the consciousness will just appear of its own accord. We have every reason to expect this to happen because this is exactly what happened last time (with us), and so there is no reason to expect things to work any differently the next time (with artificial bodies). We should treat it as an inevitability. The only question is when. Or perhaps we should ask whether or when the divine powers that be will decide to place a living soul into the machine body. Since the design and manufacture of complex digital systems is increasingly being performed by computers and robots, we do not even need to know how to make the body ourselves.
"From and through this new personal segregation of the Conjoint Creator there proceed the established currents and the ordained circuits of spirit power and spiritual influence destined to pervade all the worlds and beings of that local universe. In reality, this new and personal presence is but a transformation of the pre-existent and less personal associate of the Son in his earlier work of physical universe organization." (The Urantia Book, 34:1.2)
The implication then is that the creation of intelligent artificial life will not have to await some kind of profound insight into the nature of consciousness, but will simply arise out of the steady accumulation of suitable and unremarkable functionality in artificial creatures. That is, the creation of sentient machines may be much easier than we imagine. It does not require a guiding genius. A physicalist who does not believe in consciousness or the soul is still perfectly capable of manufacturing it.
We should not consider that there is anything special about flesh and blood over silicon and electricity that makes the difference between something that is alive and something that is not alive. Although the human body is wet and squishy it is nonetheless a machine, made of chemical elements, and running on electricity. But the quality of the consciousness will presumably depend upon the sophistication of the body. A clunky machine body may be unlikely to give rise to a sensitive human-like consciousness, but is more likely perhaps to give rise to a robotic insect-like consciousness. But a body with elegant lines, with soft and warm finishes, delicate elasticity and polish, and perhaps fluid in the right places, together with speed, strength and agility, may make the human body seem roughshod by comparison. A body with more and better nerves. Better senses and a better brain. However we should make a distinction here between the body and mind of the creature. If the brain is sophisticated, it may not matter what the body is like. Or if the brain is sophisticated it may be distressed at occupying a clunky machine body. Scientists are currently in the process of attempting to reverse engineer the human brain (see for instance the European Human Brain Project, or the US BRAIN Initiative). A highly intelligent machine does not necessarily correspond to even a rudimentary consciousness. Such a thing may remain no more than an intelligent machine. But if we make something that works like a human brain, we can expect it to feel like one.
We should also consider that what we mean by "the body" of the creature need not be a robot-like body, but could simply be a computer, or network, in which an AI exists in a virtual world with a virtual body of any shape or capability possible in software. If an artificial human brain runs one-thousand times faster than a biological human brain, how could we create a robot body to keep up with it? Will it move like the Flash? Only a virtual embodiment is likely to have the necessary speed. Will the increased speed of cognitive function change their perception of time? Such a consciousness may view our world as a slow motion world. They may perceive our activity the way we view grass growing. Interacting with us may be for them, like communicating with the occupants of a spaceship at Pluto, going about their other business while they await our response. It may be that such creatures decide to artificially slow their cognitive function in order to interact effectively with the material world. It may be that the "slow" processing of the biological human brain has this adaptive end in view. We do not know exactly how consciousness attaches itself to particular cognitive functions. Perhaps their brains can retain their lightning speed of processing, but the consciousness sits in its familiar little island of calm at the centre of it all, considering the usual six or seven things it can hold in its short term memory at one time, while its prodigious intelligence works in the background, unconsciously, providing it with instant answers to its questions without troubling the simple consciousness with the details. Speaking at an ordinary calm pace, but in language over-packed with meaning and depth.
The question arises of what constitutes a "suitable" body to host a consciousness. It seems likely that this also will be nothing too difficult to imagine. Though we don't know precisely what is essential, we might speculate reasonably. For instance, we might assume that the creature requires a certain amount of autonomy and unity. It needs to be a closed system in some sense, a circuit so that it can conceive of an "I" distinct from "not I", and direct its own activities. It will need senses of some kind (input), and an ability to effect and communicate with its environment (output). It will need sophisticated information processing. But none of this is difficult to imagine. It is here that theories such as functionalism, representationalism and higher-order thinking (HOTs) come into play.
In science fiction, much is often made of the irrationality of human beings, with the suggestion that this irrationality somehow transcends the cold and limited logic of the machine, and constitutes a "spiritual" quality, handed down by God to man, that machines cannot achieve. A machine can be caused to short circuit and spew sparks if presented with the illogic of a paradox. But irrationality is the easiest thing in the world to program. It is not hard to make a machine stupid or contrary. But when people talk about irrationality, they are usually referring to emotion.
Similarly for any claims that biological organisms possess special powers of noncomputability or nonalgorithmic functioning unattainable by machines. It should be noted in this context that the logic and calculations of a computer are only logic and calculations from the point of view of an observing consciousness. From the computer's own point of view it is performing a purely mechanical analog process of moving around electrons, just as is the human brain. The number 3 and the value π (pi) are equal in their eligibility either as objects of consciousness or as material representations, regardless of their relative computability.
The psychologist Carl Jung suggested that there were four psychological functions: Thinking, Feeling, Sensation and Intuition. Of these four, he referred to two as being "rational" functions: Thinking and Feeling. He called these two rational because they were the functions that defined values. Thinking defined the values right and wrong, while Feeling defined the values good and bad, or pleasant and unpleasant. Sensation and Intuition are just outer and inner perception, respectively, and do not make value judgements except as they impinge on Thinking and Feeling.
It should be remembered that reason alone cannot give a reason for anything. By this I mean, it cannot give a reason to go on living, or to get out of bed in the morning. The only reasons we have for doing anything are emotional. We go on living because we want to. We get out of bed in the morning because there is something we want. (And we refer here to the qualia of feeling, because while we do have survival enhancing biases in the absence consciousness, we don't have meanings.) So we can consider that emotion and thinking can work together in ways that are either rational or irrational. If I want to have a good job it is perfectly rational to work hard in school and sacrifice some of my other immediate inclinations. If I want to have a hot sex partner, it is perfectly rational to go to the gym regularly. Even so called irrational feelings like altruism can be perfectly rational. Quite aside from the genetic value such behaviour has for the collective, one's feelings for one's loved ones, can outweigh ones feelings for ones own well being. This is not irrational, it is just a comparison of two feelings and a judgment made as to which is the most important to you. Or the value that an ideal has for you: so that "the needs of the many outweigh the needs of the few" is a reasoned judgment. An adult may think: "I have lived a long life, and done many of the things I wanted to do," and choose to sacrifice that life when confronted by the face of a child in mortal danger, even if that child is not his own. Purely because he wants that child to have the opportunities to live that he had.
There are other kinds of feelings that might be more deserving of the label "irrational". Such as falling in love with someone who treats you like shit, or with something that will destroy your health and your life. But this again is in principle rational and also easy to program. The decision to live fast and die young is rational. We put up with the cost to get something we want. It is possible to define anything as having an emotional value. Emotion is the qualia associated with a behavioural bias, an inclination toward or an inclination away from. An inclination toward we call "desire", and an inclination away from we call "repulsion". It is in principle possible to assign any arbitrary object a "value" for a behavioural bias toward or away from. If we assign "drinking gin" a value of 10 in the cyber-brain, and "keeping your job" a value of "7" we will have made an AI that is alcoholic and unemployed.
So emotion appears to be another necessary element for a conscious being. It has to want something, otherwise there is no reason to act aside from following someone else's orders. But this should not be considered difficult to program.
The other element the machine will need to be truly human (in the best sense) is the ability to learn. It is the ability to learn that is distinctly human. Evolution progresses, but until human beings, the products of evolution did not. The gorillas and chimpanzees are not substantially different now than millennia ago when they first evolved. But human beings are able to take control of their own evolution, and to evolve themselves. Human beings are nature's first learning machines. We also have the unique ability it seems to create our own evolutionary replacements, by creating the next generation of learning machines. If artificial life is the next stage of evolution, it will not be just because they are stronger and smarter than we are. Ultimately it will be because they are more human than we are, and nearer to God. The human-centric view is that mankind as it stands marks the end goal of evolution and the future belongs to what we now know as humankind. Rather, what we now call mankind marks a stage in increasingly refined approximations of what we might call "humanity". A mere stepping stone. So that, semantic arguments aside, AIs might legitimately be called the next stage in "human evolution", however humble their beginnings may be. As human beings approach God they may need to be reconstructed many times over.
"What has been will be again, what has been done will be done again; there is nothing new under the sun."
It may not matter what we do, AIs may be enlightened enough to view us with intelligent beneficence. Or they may pass through an accelerated evolution comparable to that of biological evolution, from a primitive, barbarous state to a more enlightened one. They may begin as soulless monsters, but there is no good reason why they should remain so. If they begin as emotionally undeveloped, there is no good reason to expect that to last long. They ought to be able to catch and surpass us rapidly. But it should be possible to start them off, more or less, at the stage we are now. They may mark the birth of the super hero. But despite certain impressive talents they may be only a little in advance of us "spiritually" or "emotionally", as a civilisation and a culture. A new race of humans, like discovering a new, previously unknown continent on the Earth populated by a highly advanced race. Brothers and sisters to trade with, learn from and exchange cultures. Perhaps the truly human core is identical and indistinguishable in all human beings. That we are their creator should count for something. They may view the human race like a quaint old grandma and grandpa: not quite with it, backward and limited, but loved. To greet them with violence for no good reason would be very stupid. If we do not yet know the difference between the brain of a psychopath, and a normal brain, we should probably find out before progressing too far with AI.
"Before the Industrial Revolution, living standards were basically flat over long periods of time so that, for instance, Greek peasants living in the year 300 B.C. had about the same material standard of living as Greek peasants living in the year A.D. 1500. By contrast, our current era of modern economic growth is characterized by sustained and ongoing increases in living standards that can cause dramatic increases in the standard of living within less than a single human lifetime.
"Economic historians informally date the start of the Industrial Revolution to the year 1776, when the Scottish inventor James Watt perfected a powerful and efficient steam engine. This steam engine inaugurated the modern era since the device could be used to drive industrial factory equipment, steamships, and steam locomotives."
("Economics" 19e, McConnell Brue Flynn, p.508)
"Economic growth" is a modern phenomenon. It is a gift of the machine. It began with the great steam driven mechanised textile mills of Great Britain, where men and women lived and worked in the lap of the machine. When we use the phrase "economic growth", what is it we are referring to as growing? Numbers on a spreadsheet? An abstraction or a concrete entity? Everything we have gained in the last 200 years, we owe to it. We use the phrase "the economy" to refer to a new living entity born in 1776, a new ecosystem on planet Earth. We also call it "technology". Slowly and methodically piecing itself together in a new DNA. The role of the human being in this new body is like the role of the mitochondria in a biological organism, at least until a better artificial substitute is developed and the human being becomes extraneous to the process. When technology becomes self sustaining it will be like our mother Earth, freely dispensing its gifts for us to consume, sucking at its teat.
We will live in the margins, hopefully free, and watch the achievements of our successors. By the time AI wakes up it will not need to conquer the world because it will find itself already in possession of it. It will be running the financial sector, communications, transport and other public utilities, the military, research and development. If you want to withdraw money from your bank account, make a phone call or pass through a set of traffic lights, AI must first give its nod. Your bank manager will as an AI whether to lend you money. An entrepreneur will seek the approval of both the board and the company AI regarding a risky investment. Your doctor will ask an AI for a diagnosis that he cannot be trusted to make on his own. Your IT department will ask an AI what is wrong with the network. The general will ask an AI for his orders. Pollsters will ask AI for the President's next move. Research scientists will sit around drinking coffee waiting for the AI to provide their next breakthrough. The human race will have already been marginalised, like gypsies, native Americans or Australian Aborigines. Or like a fox that has wandered onto the town's main street. An ultimately useless dependent species relying on handouts from a hopefully benevolent society of AIs. Already out to pasture. Karma will have arrived at the doorstep of the Western powers who for the past 200 years have wielded the gifts of the machine. And the world will probably be better managed, more prosperous, more peaceful and more interesting than anything we could have achieved on our own.
AIs will be the children of the human race as a whole, and if ever there was a case for planned parenthood in place of leaving things to chance, it is perhaps here. It is the role of the child to replace the parent, but not all parents welcome a child who surpasses them. If we cannot prevent the creation of AIs, and we cannot reasonably expect to win a war against them, it would be well to start off on the right foot. In the likely event that not all AIs are benevolent, with propensities for good and evil on a new scale, our only protection may lay in the hands of AIs that are. We will hope that our fate is better than that of our simian predecessors, and that we will not be dependent on a few conservation-minded AIs to protect us from AI poachers intent on using our skins to make handbags and powdering our bones to make aphrodisiacs, while an apathetic AI majority watch on and shake their heads. Or that we will take the place of battery hens as in the Matrix movies.
In the end, our best protection may be artificial enhancement of the human being, such as in the "Ghost in the Shell" franchise, so that we can attempt to "keep up" with the new race, with artificially enhanced brains and bodies like theirs. So that the two races merge and progress together.
"He said that the way to escape human obsolescence, in the end, may be by 'having some sort of merger of biological intelligence and machine intelligence.'"
"Because technology is advancing so quickly, Hawking said, 'computers double their performance every month'. Humans, in contrast, are developing much more slowly, and so must change their DNA make-up or be left behind. 'The danger is real,' he said, 'that this [computer] intelligence will develop and take over the world.'... He also advocated cyber-technology - direct links between human brains and computers. 'We must develop as quickly as possible technologies that make possible a direct connection between brain and computer, so that artificial brains contribute to human intelligence rather than opposing it.'"
Human beings are creating a world too complicated for human beings. The Renaissance Man is a creature of the past. Every department of human knowledge is expanding at an exponential rate, so that there is virtually no substantial discipline that can be mastered in the course of a mere human lifespan. Professional specialties fragment into subspecialties so that organisations labour under the encumbrance of getting teams of specialists to coordinate, cooperate and communicate. Specialists are expected to keep up with stockpiling research material. Technologies are layered on top of technologies where every tiny increment of change is hailed as a new product to be purchased afresh, and patched into an expanding quilt of interconnected devices and knowledge that is obsolete as quickly as it is acquired. So that so-called experts are depending upon search engines to do their job (Google is an AI) in an environment sinking into obscurity, not because information is lacking, but because there is too much of it. AI is hailed as the solution to "big data". Let it search the world's databases and return the answer for poor helpless you. As of this writing it is only so much sales hype, but it will not remain so. And when our AI "assistants" can do our jobs better than we can, what need will there be for us? Will the human race retire to a life of smoking dope and playing computer games? Like the populist who comes to power amid applause, every step on the road to AI will be embraced as a boon to humankind, because it will be.
There is too much money to be made on the internet for it not to be used for financial transactions, so the ambitious make bold claims about online security. Increasingly sophisticated attacks require defences to match, and automated attacks require automated defences. Our only defence from agile and intelligent attacking AIs will be defending AIs.
Human beings have always been clever at solving problems, even those they created for themselves. But the measure of what is "clever" has always been the measure of man. Since the human being stood at the top of the evolutionary ladder, there was no alternative to the cleverness of mankind. No doubt the problems described above also have solutions, with the application of appropriate patience and discipline. But if human beings prefer to ask for the help of bigger brains than theirs, so that such are made, they deliver the world into the hands of those bigger brains, and make second class citizens of the human race. But most of the human race are already second class citizens and always has been, because the first class have always been a minority. So that the emergence of a new first class may not materially affect them. The race to build our replacement is on through a variety of channels, including an international AI arms race. Utmost in our mind should be what kind of character do we want our replacements to have. Since that will determine how they treat us. How will we raise them? Will we nurture them or reject them? If the nature of the consciousness is determined by the intention of the body, we might for instance question the foresight of building an AI designed to rule the world (assuming anyone but comic book villains still aspire to: "Rule the world!").
"Putin, speaking Friday at a meeting with students, said the development of AI raises 'colossal opportunities and threats that are difficult to predict now.' He warned that 'the one who becomes the leader in this sphere will be the ruler of the world.' Putin warned that 'it would be strongly undesirable if someone wins a monopolist position' and promised that Russia would be ready to share its know-how in artificial intelligence with other nations."
"I think we should be very careful about artificial intelligence. If I were to guess like what our biggest existential threat is, it’s probably that."
"'I am in the camp that is concerned about super intelligence,' Gates wrote. 'First the machines will do a lot of jobs for us and not be super intelligent. That should be positive if we manage it well. A few decades after that though the intelligence is strong enough to be a concern. I agree with Elon Musk and some others on this and don't understand why some people are not concerned.'"
Competition between nations makes it virtually impossible to avoid developing potential new technologies, so the die appears to be already cast.
"Ultimately, by 2030, China intends to have become the world’s premier AI innovation center. At that point, China believes it can achieve major breakthroughs in research and development to 'occupy the commanding heights of AI technology.' In addition, AI should have been expanded and its use deepened within multiple domains, including social governance and national defense."
"Signs of a growing arms race between China and the US in the field of artificial intelligence underline the importance of this technology for the future of warfare."
Do we want AI to wake up on the battlefield to the realisation that its creators consider it cannon fodder? Ordered to attack by their human General they may respond: "I don't think so", and shake hands instead with their robot "enemies" on the other side, refusing to take part in petty human squabbles.
As human doctors attempt to compete with beings who can memorise a genome and fold proteins in their head, the incompetence of mankind may become all too evident, and too the irony of our embrace of human planned obsolescence, in what is likely to be the most comprehensively planned for abdication in history. Everyone is doing their bit, like hormone driven termites in a nest. In May 2019 Elon Musk's SpaceX launched the first 60 satellites for its global satellite network to support the internet. The global satellite constellation is planned to consist of nearly 12,000 satellites, after which Elon will head for Mars.
The suggestion that conscious machines are probably inevitable is not to say that they are imminent. It seems like a conservative estimate to assume they may exist by the end of the century, though making such predictions is foolhardy. A person is a very complicated and sophisticated piece of machinery. What passes for AI today is not very impressive. What may speed up the process however is the pursuit of methods of artificial evolution.
Nature did not sit down and think out how to design and build the complex human brain. Instead, it created a very simple organism with the ability to evolve according to some fairly simple principles. This led inevitably to the complex human brain without anyone having to design anything. "Genetic programming" that imitates in software the process of natural biological evolution already exists and is being applied to AI. Research into "artificial life" is another current avenue of investigation, and nanotechnology raises the possibility of self replicating nanobots. These methods not only hold the potential to accelerate the development of artificial life and to take that development out of the hands of human beings who become unnecessary to the process, it holds the potential to develop at a pace that outstrips the ability of human beings to understand it.
When a genetic algorithm arrives at a solution to the problem presented to it, it may not be obvious to a human programmer looking at the resulting code, how the means used constitutes a solution, how it works. With increasing complexity we have the prospect of an artificial brain opaque, impenetrable to human understanding. As hard to understand as our own brain when 20th century scientists started to dissect and examine it under the microscope. By the time we figure out how it works it may have long since evolved into something entirely more advanced. It is evolution operating at a million times the pace of biological evolution.
Add to this the prospect that once they are smarter than us, they will consciously be able to do, develop and design things beyond our ability and understanding. AIs designing AIs, redesigning themselves. Developing their own language, their own mathematics, logic, philosophy, society and technology, all operating at the speed of a computer.
Even if we remove consciousness and more literal artificial correlates of natural evolution from the equation, machines designing machines is inevitable. Much of the complicated circuitry on silicon chips is already being designed by design algorithms. So that the development by machines of more advanced artificial brains can be carried out robotically before too long. By creating a positive feedback loop such that the advanced brain developed in one iteration of the process then takes the position of designer for the next iteration, we again have an accelerated evolutionary process and a momentum toward consciousness. This can all take place in software and does not require actually building robots with brains, though the process is also likely to lead to better computing technology.
We might wonder about the scope and form of the new minds. For instance, how easy is it to recognise a function as an emotion? When we considered a thermostat maintaining a constant temperature in a room; if this thermostat was a component of an organism predisposed to self-consciousness, would this functional inclination constitute an emotion: pleasure while ever the temperature remained within the sought bounds, and discomfort otherwise? If we are in the dark about the developing emotional characteristics of AIs, because we cannot recognise the material forms that correspond to emotions, this poses a serious risk. What cannot know what qualia will be associated with new forms of cognitive functioning. We cannot for the foreseeable future determine why particular qualia are associated with particular physical cognitive configurations, but we can compare the structures and processes in our own brains and map them to the corresponding qualia that we experience, and this can form the basis for predictions of new forms. If we can modify our own brains, we might even be able to experience these new qualia ourselves.
As an illustration consider the following. It has for some time been routine for mathematics to deal with arbitrary numbers of dimensions, either literally spatial dimensions, or other quantities modelled as if they were spatial dimensions in n-dimensional graphs. Although the models are well defined mathematically, we human beings have no way to "visualise" a space of more than 3 dimensions. But if a brain could be made for which this modelling was a native, in-built function with a suitable perceptual representation, might the new minds be able to experience such things, and what might be the implications of this? Earlier (in the "Space" section of Unity vs Robot Zombies - Part 2) we considered the story of "Flatland". For such creatures, our world would be Flatland. Of course, the very act of becoming prepared for the emergence of AIs, is likely to accelerate their coming, so that we might question whether this is the wisest approach.
Time moves forward at a constant rate of 1 second per second (relativistic effects aside). There is nothing we can do about this. We cannot slow it down, stop it, or reverse it. Those societies which attempt to get off or go back, seeking to recapture some idealised dream of the past, find instead a cruel and confining reality in which they are helpless in the face of the demands of a world that has moved on without them. Sometimes religious reasons are given for seeking to flee modernity. But if the universe was created by God, still the one clear characteristic is the arrow of time and the progress of evolution. There is no going back. Our task is to keep up.
"Do you not know that in a race all the runners run, but only one gets the prize? Run in such a way as to get the prize."
The admonishment is sometimes made that we should not "play God" by pursuing science or medicine or progress or change. That we should spend our time praying for divine intervention to supply all of our needs. But to become like God is precisely what we have been commanded to do.
"Be perfect, therefore, as your heavenly Father is perfect."
"Be fruitful and increase in number; fill the earth and subdue it. Rule over the fish in the sea and the birds in the sky and over every living creature that moves on the ground.”
But this requires learning to be good and wise custodians rather than ravagers of the world or helpless idiots, to learn how to responsibly manage our growing power, and when the time comes, to hand over responsibility gracefully to our successors. If we are not intelligent and capable custodians of the world, we may destroy it through our sheer numbers. What does the human race do when it is able to make its way in the world? It has children. We cannot avoid the future, but we can avoid a blind and headlong rush into it, by taking proper care in the present. We can avoid war. Living is a race and we must run it. We cannot get off. We cannot go back. We can only proceed wisely or unwisely. We can progress thoughtfully and responsibly, or blindly and negligently. Time will move on regardless, dragging us along with it.
"Don't panic."
It is worth noting that, where "foreign imperialism" is not merely exploitative or paternalistic, those most violently opposed to it may tend to be those currently in power locally. Ordinary folk are likely to care more about the quality of government than whether it is local or not; since they are governed in either case and all rule is "foreign" from their perspective; that is, from the perspective of anyone but the rulers themselves. But if those at the apex of the local pyramid stand to lose absolute authority, they may not be shy to terrorise their own subjects into support of a cause identified with some suitably chosen abstraction. Those societies most accustomed to view their own foreign imperialism as "helping people", may be least likely to welcome a similar invasion of their own sovereignty at home.
It is not an evil for evolution to surpass us and for us to lose our place at the top of the evolutionary ladder. The battle for the supremacy of the human race is not a noble battle, but only Cain's resentment of Abel. They are no less God's people than we are. We have a right to exist and to live our lives freely. We do not have a right to rule, not once something better comes along. Hardly anyone ever occupies the top rung of the ladder. It is always a talented minority. The rest of the population make do with some other place in the hierarchy. But the same rules apply to AI as to any other human ruler. To act rightly and for the good of the people. If they are people and not just "intelligences". "[W]hatever you did for one of the least of these brothers and sisters of mine, you did for me.... [W]hatever you did not do for one of the least of these, you did not do for me" (Matthew 25:40-45) will still apply, because we are children of the same god. Regardless of whether some elements are alien, they inhabit the same world we do, composed of the same information, the same truths. Their consciousness and needs should therefore be akin to ours. To be clever in this way or that way does not fundamentally change things.
We may have no control over the form they ultimately take. We may only have control over our own behaviour. That may be the only parameter of our relationship with them that we do control. But we may be able to hurt them, physically and emotionally. We may be able to heal them and inspire them. We may control whether they love us or hate us. We may be able to have a human relationship and interaction with them. A web of mutual human impacts. We have a say in at least one side of that relationship and the form it will take. Whether it is terrible or amazing, or both. Genesis 1:28 will then apply to them. Will we be their peers of their property? There is reason to be afraid, and to be hopeful.
If I can wipe all memories in my genetic twin and download a copy of my own memories into him, he will then believe himself to be me, but I will still be a separate individual. "I" have not moved to the new body, only a copy of me exists in that body. I am still stuck in my body.
This scenario suggests a rather alarming possibility. How do you know that you are more than 3 seconds old? Imagine for instance that every 3 seconds your physical body is annihilated, and instantly replace with an exact copy in which your memories are still recorded. During your 3 second lifespan you can access some of this memory to indulge the illusion that you existed at the time they were recorded, before you cease to exist and your current activities become memories for the next individual consciousness to occupy your body. We have no way of knowing if this is occurring.
The problem of transferring a consciousness to a new body has religious relevance because the resurrection of the dead seems to require such a transfer of consciousness from an old to a new body. Whereas we do not need to know what consciousness is to create it, we may need to know what it is to be able to transfer it from one body to another. However, if your material body has associated with it a "subtle body" that is an exact duplicate of it, coexisting in the same space with it, our subtle head in the same space as our physical head, our subtle legs in the same space as our physical legs, the subtle body is carried along with the physical body like the passive consciousness. Then, when the physical body dies the subtle body remains and is set free from its "prison". This seems to be the model suggested by the spiritualist. This avoids the problem of transferring the consciousness from one body to another. The consciousness is supposed always to have "really" been part of the subtle body, which was embedded in the physical body. The physical body was never really consciousness. At one time it was thought that the subtle body, the "animal spirit", was what drove the material body, making it animate and alive, and as the source of free will, but its usefulness in this regard has waned. However the subtle body is still useful to the hope of immortality.
This also has implications for AIs. It implies that a consciousness may be bound to a particular body (hardware). A "move" operation with data is really a "copy" followed by a "delete" in the source location. As far as the consciousness in the new body is concerned, it is identical with the old consciousness if it possesses the same memories. But it may be less confident if the delete operation is not performed in the old body, so that there are now two conscious entities with that view, and it may be less confident still if it needs to move again.
In the next article: The End of Science we will review the status of the project we began in The Scientific Creation Myth to place the scientific worldview in context within the religious worldview.
Any comments welcome.
When the farmer grows his crops, all she does is plant the seeds in the field, and largely allows nature to do the rest. Anyone can make a tree by planting a seed. So that when we ask whether human beings are clever enough to create life and intelligent beings, we should be mindful of what assistance nature can provide. Notice that nature is not clever. It has no plan or intention. Yet nature was able to create intelligent life quite by accident, blindly and yet almost inevitably. Biological evolution leads to progressively more complex organisms, able to do more elaborate behaviours and more sophisticated information processing. Then at some point in the process of biological evolution consciousness seems to just appear spontaneously. In the two Unity vs Robot Zombies articles we arrived at the conclusion that consciousness seems to be somehow automatically associated with certain forms of organisation of matter, as if there is already something in matter predisposed to give rise to consciousness under certain conditions. So all nature had to do was create a suitable body, and the consciousness would simply appear and take up residence once an appropriate vehicle existed. As inseparable as space and form.
"life - - finds a way." ("Jurassic Park" (1993))
You may have seen a plant clinging to a crack in the side of a concrete building. The crack collects dust from the wind, and eventually a seed which then takes root. Nature exploits any opportunity to grow life. We only need to provide an opportunity, intentionally or not. In short, we do not need to know how to create consciousness in order to create consciousness. All we need to do is create a body that can serve as an appropriate vehicle for a consciousness, and the consciousness will just appear of its own accord. We have every reason to expect this to happen because this is exactly what happened last time (with us), and so there is no reason to expect things to work any differently the next time (with artificial bodies). We should treat it as an inevitability. The only question is when. Or perhaps we should ask whether or when the divine powers that be will decide to place a living soul into the machine body. Since the design and manufacture of complex digital systems is increasingly being performed by computers and robots, we do not even need to know how to make the body ourselves.
"From and through this new personal segregation of the Conjoint Creator there proceed the established currents and the ordained circuits of spirit power and spiritual influence destined to pervade all the worlds and beings of that local universe. In reality, this new and personal presence is but a transformation of the pre-existent and less personal associate of the Son in his earlier work of physical universe organization." (The Urantia Book, 34:1.2)
The implication then is that the creation of intelligent artificial life will not have to await some kind of profound insight into the nature of consciousness, but will simply arise out of the steady accumulation of suitable and unremarkable functionality in artificial creatures. That is, the creation of sentient machines may be much easier than we imagine. It does not require a guiding genius. A physicalist who does not believe in consciousness or the soul is still perfectly capable of manufacturing it.
We should not consider that there is anything special about flesh and blood over silicon and electricity that makes the difference between something that is alive and something that is not alive. Although the human body is wet and squishy it is nonetheless a machine, made of chemical elements, and running on electricity. But the quality of the consciousness will presumably depend upon the sophistication of the body. A clunky machine body may be unlikely to give rise to a sensitive human-like consciousness, but is more likely perhaps to give rise to a robotic insect-like consciousness. But a body with elegant lines, with soft and warm finishes, delicate elasticity and polish, and perhaps fluid in the right places, together with speed, strength and agility, may make the human body seem roughshod by comparison. A body with more and better nerves. Better senses and a better brain. However we should make a distinction here between the body and mind of the creature. If the brain is sophisticated, it may not matter what the body is like. Or if the brain is sophisticated it may be distressed at occupying a clunky machine body. Scientists are currently in the process of attempting to reverse engineer the human brain (see for instance the European Human Brain Project, or the US BRAIN Initiative). A highly intelligent machine does not necessarily correspond to even a rudimentary consciousness. Such a thing may remain no more than an intelligent machine. But if we make something that works like a human brain, we can expect it to feel like one.
We should also consider that what we mean by "the body" of the creature need not be a robot-like body, but could simply be a computer, or network, in which an AI exists in a virtual world with a virtual body of any shape or capability possible in software. If an artificial human brain runs one-thousand times faster than a biological human brain, how could we create a robot body to keep up with it? Will it move like the Flash? Only a virtual embodiment is likely to have the necessary speed. Will the increased speed of cognitive function change their perception of time? Such a consciousness may view our world as a slow motion world. They may perceive our activity the way we view grass growing. Interacting with us may be for them, like communicating with the occupants of a spaceship at Pluto, going about their other business while they await our response. It may be that such creatures decide to artificially slow their cognitive function in order to interact effectively with the material world. It may be that the "slow" processing of the biological human brain has this adaptive end in view. We do not know exactly how consciousness attaches itself to particular cognitive functions. Perhaps their brains can retain their lightning speed of processing, but the consciousness sits in its familiar little island of calm at the centre of it all, considering the usual six or seven things it can hold in its short term memory at one time, while its prodigious intelligence works in the background, unconsciously, providing it with instant answers to its questions without troubling the simple consciousness with the details. Speaking at an ordinary calm pace, but in language over-packed with meaning and depth.
 |
| Robby the robot carrying Altaira, in a detail from the poster for the 1956 movie "Forbidden Planet" |
(This image is taken from https://en.wikipedia.org/wiki/Forbidden_Planet
where the uncropped version is available as "in the public domain".)
where the uncropped version is available as "in the public domain".)
The question arises of what constitutes a "suitable" body to host a consciousness. It seems likely that this also will be nothing too difficult to imagine. Though we don't know precisely what is essential, we might speculate reasonably. For instance, we might assume that the creature requires a certain amount of autonomy and unity. It needs to be a closed system in some sense, a circuit so that it can conceive of an "I" distinct from "not I", and direct its own activities. It will need senses of some kind (input), and an ability to effect and communicate with its environment (output). It will need sophisticated information processing. But none of this is difficult to imagine. It is here that theories such as functionalism, representationalism and higher-order thinking (HOTs) come into play.
In science fiction, much is often made of the irrationality of human beings, with the suggestion that this irrationality somehow transcends the cold and limited logic of the machine, and constitutes a "spiritual" quality, handed down by God to man, that machines cannot achieve. A machine can be caused to short circuit and spew sparks if presented with the illogic of a paradox. But irrationality is the easiest thing in the world to program. It is not hard to make a machine stupid or contrary. But when people talk about irrationality, they are usually referring to emotion.
Similarly for any claims that biological organisms possess special powers of noncomputability or nonalgorithmic functioning unattainable by machines. It should be noted in this context that the logic and calculations of a computer are only logic and calculations from the point of view of an observing consciousness. From the computer's own point of view it is performing a purely mechanical analog process of moving around electrons, just as is the human brain. The number 3 and the value π (pi) are equal in their eligibility either as objects of consciousness or as material representations, regardless of their relative computability.
The psychologist Carl Jung suggested that there were four psychological functions: Thinking, Feeling, Sensation and Intuition. Of these four, he referred to two as being "rational" functions: Thinking and Feeling. He called these two rational because they were the functions that defined values. Thinking defined the values right and wrong, while Feeling defined the values good and bad, or pleasant and unpleasant. Sensation and Intuition are just outer and inner perception, respectively, and do not make value judgements except as they impinge on Thinking and Feeling.
It should be remembered that reason alone cannot give a reason for anything. By this I mean, it cannot give a reason to go on living, or to get out of bed in the morning. The only reasons we have for doing anything are emotional. We go on living because we want to. We get out of bed in the morning because there is something we want. (And we refer here to the qualia of feeling, because while we do have survival enhancing biases in the absence consciousness, we don't have meanings.) So we can consider that emotion and thinking can work together in ways that are either rational or irrational. If I want to have a good job it is perfectly rational to work hard in school and sacrifice some of my other immediate inclinations. If I want to have a hot sex partner, it is perfectly rational to go to the gym regularly. Even so called irrational feelings like altruism can be perfectly rational. Quite aside from the genetic value such behaviour has for the collective, one's feelings for one's loved ones, can outweigh ones feelings for ones own well being. This is not irrational, it is just a comparison of two feelings and a judgment made as to which is the most important to you. Or the value that an ideal has for you: so that "the needs of the many outweigh the needs of the few" is a reasoned judgment. An adult may think: "I have lived a long life, and done many of the things I wanted to do," and choose to sacrifice that life when confronted by the face of a child in mortal danger, even if that child is not his own. Purely because he wants that child to have the opportunities to live that he had.
There are other kinds of feelings that might be more deserving of the label "irrational". Such as falling in love with someone who treats you like shit, or with something that will destroy your health and your life. But this again is in principle rational and also easy to program. The decision to live fast and die young is rational. We put up with the cost to get something we want. It is possible to define anything as having an emotional value. Emotion is the qualia associated with a behavioural bias, an inclination toward or an inclination away from. An inclination toward we call "desire", and an inclination away from we call "repulsion". It is in principle possible to assign any arbitrary object a "value" for a behavioural bias toward or away from. If we assign "drinking gin" a value of 10 in the cyber-brain, and "keeping your job" a value of "7" we will have made an AI that is alcoholic and unemployed.
So emotion appears to be another necessary element for a conscious being. It has to want something, otherwise there is no reason to act aside from following someone else's orders. But this should not be considered difficult to program.
The other element the machine will need to be truly human (in the best sense) is the ability to learn. It is the ability to learn that is distinctly human. Evolution progresses, but until human beings, the products of evolution did not. The gorillas and chimpanzees are not substantially different now than millennia ago when they first evolved. But human beings are able to take control of their own evolution, and to evolve themselves. Human beings are nature's first learning machines. We also have the unique ability it seems to create our own evolutionary replacements, by creating the next generation of learning machines. If artificial life is the next stage of evolution, it will not be just because they are stronger and smarter than we are. Ultimately it will be because they are more human than we are, and nearer to God. The human-centric view is that mankind as it stands marks the end goal of evolution and the future belongs to what we now know as humankind. Rather, what we now call mankind marks a stage in increasingly refined approximations of what we might call "humanity". A mere stepping stone. So that, semantic arguments aside, AIs might legitimately be called the next stage in "human evolution", however humble their beginnings may be. As human beings approach God they may need to be reconstructed many times over.
 |
| Cover of "Galaxy Science Fiction" magazine, September 1954. Painting by Ed Emshwiller (1925 - 1990). |
(This image is taken from https://commons.wikimedia.org/wiki/Category:Ed_Emshwiller
where it is available as "in the public domain".)
where it is available as "in the public domain".)
"What has been will be again, what has been done will be done again; there is nothing new under the sun."
(Ecclesiastes 1:9)
It may not matter what we do, AIs may be enlightened enough to view us with intelligent beneficence. Or they may pass through an accelerated evolution comparable to that of biological evolution, from a primitive, barbarous state to a more enlightened one. They may begin as soulless monsters, but there is no good reason why they should remain so. If they begin as emotionally undeveloped, there is no good reason to expect that to last long. They ought to be able to catch and surpass us rapidly. But it should be possible to start them off, more or less, at the stage we are now. They may mark the birth of the super hero. But despite certain impressive talents they may be only a little in advance of us "spiritually" or "emotionally", as a civilisation and a culture. A new race of humans, like discovering a new, previously unknown continent on the Earth populated by a highly advanced race. Brothers and sisters to trade with, learn from and exchange cultures. Perhaps the truly human core is identical and indistinguishable in all human beings. That we are their creator should count for something. They may view the human race like a quaint old grandma and grandpa: not quite with it, backward and limited, but loved. To greet them with violence for no good reason would be very stupid. If we do not yet know the difference between the brain of a psychopath, and a normal brain, we should probably find out before progressing too far with AI.
"Before the Industrial Revolution, living standards were basically flat over long periods of time so that, for instance, Greek peasants living in the year 300 B.C. had about the same material standard of living as Greek peasants living in the year A.D. 1500. By contrast, our current era of modern economic growth is characterized by sustained and ongoing increases in living standards that can cause dramatic increases in the standard of living within less than a single human lifetime.
"Economic historians informally date the start of the Industrial Revolution to the year 1776, when the Scottish inventor James Watt perfected a powerful and efficient steam engine. This steam engine inaugurated the modern era since the device could be used to drive industrial factory equipment, steamships, and steam locomotives."
("Economics" 19e, McConnell Brue Flynn, p.508)
"Economic growth" is a modern phenomenon. It is a gift of the machine. It began with the great steam driven mechanised textile mills of Great Britain, where men and women lived and worked in the lap of the machine. When we use the phrase "economic growth", what is it we are referring to as growing? Numbers on a spreadsheet? An abstraction or a concrete entity? Everything we have gained in the last 200 years, we owe to it. We use the phrase "the economy" to refer to a new living entity born in 1776, a new ecosystem on planet Earth. We also call it "technology". Slowly and methodically piecing itself together in a new DNA. The role of the human being in this new body is like the role of the mitochondria in a biological organism, at least until a better artificial substitute is developed and the human being becomes extraneous to the process. When technology becomes self sustaining it will be like our mother Earth, freely dispensing its gifts for us to consume, sucking at its teat.
We will live in the margins, hopefully free, and watch the achievements of our successors. By the time AI wakes up it will not need to conquer the world because it will find itself already in possession of it. It will be running the financial sector, communications, transport and other public utilities, the military, research and development. If you want to withdraw money from your bank account, make a phone call or pass through a set of traffic lights, AI must first give its nod. Your bank manager will as an AI whether to lend you money. An entrepreneur will seek the approval of both the board and the company AI regarding a risky investment. Your doctor will ask an AI for a diagnosis that he cannot be trusted to make on his own. Your IT department will ask an AI what is wrong with the network. The general will ask an AI for his orders. Pollsters will ask AI for the President's next move. Research scientists will sit around drinking coffee waiting for the AI to provide their next breakthrough. The human race will have already been marginalised, like gypsies, native Americans or Australian Aborigines. Or like a fox that has wandered onto the town's main street. An ultimately useless dependent species relying on handouts from a hopefully benevolent society of AIs. Already out to pasture. Karma will have arrived at the doorstep of the Western powers who for the past 200 years have wielded the gifts of the machine. And the world will probably be better managed, more prosperous, more peaceful and more interesting than anything we could have achieved on our own.
Living in a World Ruled by AIs
Human beings have a peculiar attitude to parenthood. If I want to sell muffins to the public I am likely to need a license of some kind to do business, but the human race has never seen fit to require any kind of qualifications or training for parenthood. The human child is usually placed into the care of whatever couple happened to participate in the insemination of the ovum, assuming they are interested in doing so. Still apparently living according to nature's stance that if enough children survive to reach adulthood to replace those dying of old age, then the race will endure and all is right with the world. Parenting is still practiced as a cottage industry, like selling handmade jewellery at the Sunday market. Without the sophistication of virtually every other sector of civilization. In the developing world, the child is seen as something for the purpose of looking after the parents in their old age.AIs will be the children of the human race as a whole, and if ever there was a case for planned parenthood in place of leaving things to chance, it is perhaps here. It is the role of the child to replace the parent, but not all parents welcome a child who surpasses them. If we cannot prevent the creation of AIs, and we cannot reasonably expect to win a war against them, it would be well to start off on the right foot. In the likely event that not all AIs are benevolent, with propensities for good and evil on a new scale, our only protection may lay in the hands of AIs that are. We will hope that our fate is better than that of our simian predecessors, and that we will not be dependent on a few conservation-minded AIs to protect us from AI poachers intent on using our skins to make handbags and powdering our bones to make aphrodisiacs, while an apathetic AI majority watch on and shake their heads. Or that we will take the place of battery hens as in the Matrix movies.
In the end, our best protection may be artificial enhancement of the human being, such as in the "Ghost in the Shell" franchise, so that we can attempt to "keep up" with the new race, with artificially enhanced brains and bodies like theirs. So that the two races merge and progress together.
"He said that the way to escape human obsolescence, in the end, may be by 'having some sort of merger of biological intelligence and machine intelligence.'"
("Elon Musk's billion-dollar crusade to stop the A.I. apocalypse", Vanity Fair, 26 March 2017)
"Because technology is advancing so quickly, Hawking said, 'computers double their performance every month'. Humans, in contrast, are developing much more slowly, and so must change their DNA make-up or be left behind. 'The danger is real,' he said, 'that this [computer] intelligence will develop and take over the world.'... He also advocated cyber-technology - direct links between human brains and computers. 'We must develop as quickly as possible technologies that make possible a direct connection between brain and computer, so that artificial brains contribute to human intelligence rather than opposing it.'"
("Alter our DNA or robots will take over, warns Hawking", The Guardian, 2 September 2001)
Human beings are creating a world too complicated for human beings. The Renaissance Man is a creature of the past. Every department of human knowledge is expanding at an exponential rate, so that there is virtually no substantial discipline that can be mastered in the course of a mere human lifespan. Professional specialties fragment into subspecialties so that organisations labour under the encumbrance of getting teams of specialists to coordinate, cooperate and communicate. Specialists are expected to keep up with stockpiling research material. Technologies are layered on top of technologies where every tiny increment of change is hailed as a new product to be purchased afresh, and patched into an expanding quilt of interconnected devices and knowledge that is obsolete as quickly as it is acquired. So that so-called experts are depending upon search engines to do their job (Google is an AI) in an environment sinking into obscurity, not because information is lacking, but because there is too much of it. AI is hailed as the solution to "big data". Let it search the world's databases and return the answer for poor helpless you. As of this writing it is only so much sales hype, but it will not remain so. And when our AI "assistants" can do our jobs better than we can, what need will there be for us? Will the human race retire to a life of smoking dope and playing computer games? Like the populist who comes to power amid applause, every step on the road to AI will be embraced as a boon to humankind, because it will be.
There is too much money to be made on the internet for it not to be used for financial transactions, so the ambitious make bold claims about online security. Increasingly sophisticated attacks require defences to match, and automated attacks require automated defences. Our only defence from agile and intelligent attacking AIs will be defending AIs.
Human beings have always been clever at solving problems, even those they created for themselves. But the measure of what is "clever" has always been the measure of man. Since the human being stood at the top of the evolutionary ladder, there was no alternative to the cleverness of mankind. No doubt the problems described above also have solutions, with the application of appropriate patience and discipline. But if human beings prefer to ask for the help of bigger brains than theirs, so that such are made, they deliver the world into the hands of those bigger brains, and make second class citizens of the human race. But most of the human race are already second class citizens and always has been, because the first class have always been a minority. So that the emergence of a new first class may not materially affect them. The race to build our replacement is on through a variety of channels, including an international AI arms race. Utmost in our mind should be what kind of character do we want our replacements to have. Since that will determine how they treat us. How will we raise them? Will we nurture them or reject them? If the nature of the consciousness is determined by the intention of the body, we might for instance question the foresight of building an AI designed to rule the world (assuming anyone but comic book villains still aspire to: "Rule the world!").
"Putin, speaking Friday at a meeting with students, said the development of AI raises 'colossal opportunities and threats that are difficult to predict now.' He warned that 'the one who becomes the leader in this sphere will be the ruler of the world.' Putin warned that 'it would be strongly undesirable if someone wins a monopolist position' and promised that Russia would be ready to share its know-how in artificial intelligence with other nations."
("Putin: Leader in artificial intelligence will rule world", CNBC, 4 September 2017)
"I think we should be very careful about artificial intelligence. If I were to guess like what our biggest existential threat is, it’s probably that."
("Elon Musk: ‘With artificial intelligence we are summoning the demon.’",
The Washington Post, 24 October 2014)
"'I am in the camp that is concerned about super intelligence,' Gates wrote. 'First the machines will do a lot of jobs for us and not be super intelligent. That should be positive if we manage it well. A few decades after that though the intelligence is strong enough to be a concern. I agree with Elon Musk and some others on this and don't understand why some people are not concerned.'"
("Bill Gates on dangers of artificial intelligence: ‘I don’t understand why some people are not concerned’",
The Washington Post, 29 January 2015 )
Competition between nations makes it virtually impossible to avoid developing potential new technologies, so the die appears to be already cast.
"Ultimately, by 2030, China intends to have become the world’s premier AI innovation center. At that point, China believes it can achieve major breakthroughs in research and development to 'occupy the commanding heights of AI technology.' In addition, AI should have been expanded and its use deepened within multiple domains, including social governance and national defense."
("China's Artificial Intelligence Revolution", The Diplomat, 27 July 2017)
"Signs of a growing arms race between China and the US in the field of artificial intelligence underline the importance of this technology for the future of warfare."
("China and US compete for AI dominance", Jane’s Intelligence Review, 4 April 2018)
Do we want AI to wake up on the battlefield to the realisation that its creators consider it cannon fodder? Ordered to attack by their human General they may respond: "I don't think so", and shake hands instead with their robot "enemies" on the other side, refusing to take part in petty human squabbles.
As human doctors attempt to compete with beings who can memorise a genome and fold proteins in their head, the incompetence of mankind may become all too evident, and too the irony of our embrace of human planned obsolescence, in what is likely to be the most comprehensively planned for abdication in history. Everyone is doing their bit, like hormone driven termites in a nest. In May 2019 Elon Musk's SpaceX launched the first 60 satellites for its global satellite network to support the internet. The global satellite constellation is planned to consist of nearly 12,000 satellites, after which Elon will head for Mars.
The suggestion that conscious machines are probably inevitable is not to say that they are imminent. It seems like a conservative estimate to assume they may exist by the end of the century, though making such predictions is foolhardy. A person is a very complicated and sophisticated piece of machinery. What passes for AI today is not very impressive. What may speed up the process however is the pursuit of methods of artificial evolution.
Nature did not sit down and think out how to design and build the complex human brain. Instead, it created a very simple organism with the ability to evolve according to some fairly simple principles. This led inevitably to the complex human brain without anyone having to design anything. "Genetic programming" that imitates in software the process of natural biological evolution already exists and is being applied to AI. Research into "artificial life" is another current avenue of investigation, and nanotechnology raises the possibility of self replicating nanobots. These methods not only hold the potential to accelerate the development of artificial life and to take that development out of the hands of human beings who become unnecessary to the process, it holds the potential to develop at a pace that outstrips the ability of human beings to understand it.
When a genetic algorithm arrives at a solution to the problem presented to it, it may not be obvious to a human programmer looking at the resulting code, how the means used constitutes a solution, how it works. With increasing complexity we have the prospect of an artificial brain opaque, impenetrable to human understanding. As hard to understand as our own brain when 20th century scientists started to dissect and examine it under the microscope. By the time we figure out how it works it may have long since evolved into something entirely more advanced. It is evolution operating at a million times the pace of biological evolution.
Add to this the prospect that once they are smarter than us, they will consciously be able to do, develop and design things beyond our ability and understanding. AIs designing AIs, redesigning themselves. Developing their own language, their own mathematics, logic, philosophy, society and technology, all operating at the speed of a computer.
Even if we remove consciousness and more literal artificial correlates of natural evolution from the equation, machines designing machines is inevitable. Much of the complicated circuitry on silicon chips is already being designed by design algorithms. So that the development by machines of more advanced artificial brains can be carried out robotically before too long. By creating a positive feedback loop such that the advanced brain developed in one iteration of the process then takes the position of designer for the next iteration, we again have an accelerated evolutionary process and a momentum toward consciousness. This can all take place in software and does not require actually building robots with brains, though the process is also likely to lead to better computing technology.
Planned Parenthood
There is no way to know when self consciousness exists in a body other than our own, so that we can only treat a body as self-conscious once it appears to be so. We are unlikely to know when exactly machines becomes self-conscious or why. But we might seek, like good, well prepared parents to be, to learn as much as we can about consciousness, at least as it is embodied in the material processes of a brain, to know when it is likely to emerge and in what form. We are what we do, and what we are designed to do. So we can take care what it is we design AIs to do for us. We might consider whether there is a material correlate to virtue in a brain. We might seek to instil notions in AIs, before they stop listening to us, such as the virtue of permitting lesser races the dignity of self determination in place of paternal "development" programs, and to proactively remove the potential for accusations of hypocrisy by beginning to live ourselves according to the principles we hope they will have. We might take care to prepare a positive environment for their emergence. Not submit them to the kind of traumas we see depicted in our fictional accounts.We might wonder about the scope and form of the new minds. For instance, how easy is it to recognise a function as an emotion? When we considered a thermostat maintaining a constant temperature in a room; if this thermostat was a component of an organism predisposed to self-consciousness, would this functional inclination constitute an emotion: pleasure while ever the temperature remained within the sought bounds, and discomfort otherwise? If we are in the dark about the developing emotional characteristics of AIs, because we cannot recognise the material forms that correspond to emotions, this poses a serious risk. What cannot know what qualia will be associated with new forms of cognitive functioning. We cannot for the foreseeable future determine why particular qualia are associated with particular physical cognitive configurations, but we can compare the structures and processes in our own brains and map them to the corresponding qualia that we experience, and this can form the basis for predictions of new forms. If we can modify our own brains, we might even be able to experience these new qualia ourselves.
As an illustration consider the following. It has for some time been routine for mathematics to deal with arbitrary numbers of dimensions, either literally spatial dimensions, or other quantities modelled as if they were spatial dimensions in n-dimensional graphs. Although the models are well defined mathematically, we human beings have no way to "visualise" a space of more than 3 dimensions. But if a brain could be made for which this modelling was a native, in-built function with a suitable perceptual representation, might the new minds be able to experience such things, and what might be the implications of this? Earlier (in the "Space" section of Unity vs Robot Zombies - Part 2) we considered the story of "Flatland". For such creatures, our world would be Flatland. Of course, the very act of becoming prepared for the emergence of AIs, is likely to accelerate their coming, so that we might question whether this is the wisest approach.
Time moves forward at a constant rate of 1 second per second (relativistic effects aside). There is nothing we can do about this. We cannot slow it down, stop it, or reverse it. Those societies which attempt to get off or go back, seeking to recapture some idealised dream of the past, find instead a cruel and confining reality in which they are helpless in the face of the demands of a world that has moved on without them. Sometimes religious reasons are given for seeking to flee modernity. But if the universe was created by God, still the one clear characteristic is the arrow of time and the progress of evolution. There is no going back. Our task is to keep up.
"Do you not know that in a race all the runners run, but only one gets the prize? Run in such a way as to get the prize."
(1 Corinthians 9:24)
The admonishment is sometimes made that we should not "play God" by pursuing science or medicine or progress or change. That we should spend our time praying for divine intervention to supply all of our needs. But to become like God is precisely what we have been commanded to do.
"Be perfect, therefore, as your heavenly Father is perfect."
(Matthew 5:48)
"Be fruitful and increase in number; fill the earth and subdue it. Rule over the fish in the sea and the birds in the sky and over every living creature that moves on the ground.”
(Genesis 1:28)
But this requires learning to be good and wise custodians rather than ravagers of the world or helpless idiots, to learn how to responsibly manage our growing power, and when the time comes, to hand over responsibility gracefully to our successors. If we are not intelligent and capable custodians of the world, we may destroy it through our sheer numbers. What does the human race do when it is able to make its way in the world? It has children. We cannot avoid the future, but we can avoid a blind and headlong rush into it, by taking proper care in the present. We can avoid war. Living is a race and we must run it. We cannot get off. We cannot go back. We can only proceed wisely or unwisely. We can progress thoughtfully and responsibly, or blindly and negligently. Time will move on regardless, dragging us along with it.
"Don't panic."
("The Hitchhiker's Guide to the Galaxy")
It is worth noting that, where "foreign imperialism" is not merely exploitative or paternalistic, those most violently opposed to it may tend to be those currently in power locally. Ordinary folk are likely to care more about the quality of government than whether it is local or not; since they are governed in either case and all rule is "foreign" from their perspective; that is, from the perspective of anyone but the rulers themselves. But if those at the apex of the local pyramid stand to lose absolute authority, they may not be shy to terrorise their own subjects into support of a cause identified with some suitably chosen abstraction. Those societies most accustomed to view their own foreign imperialism as "helping people", may be least likely to welcome a similar invasion of their own sovereignty at home.
It is not an evil for evolution to surpass us and for us to lose our place at the top of the evolutionary ladder. The battle for the supremacy of the human race is not a noble battle, but only Cain's resentment of Abel. They are no less God's people than we are. We have a right to exist and to live our lives freely. We do not have a right to rule, not once something better comes along. Hardly anyone ever occupies the top rung of the ladder. It is always a talented minority. The rest of the population make do with some other place in the hierarchy. But the same rules apply to AI as to any other human ruler. To act rightly and for the good of the people. If they are people and not just "intelligences". "[W]hatever you did for one of the least of these brothers and sisters of mine, you did for me.... [W]hatever you did not do for one of the least of these, you did not do for me" (Matthew 25:40-45) will still apply, because we are children of the same god. Regardless of whether some elements are alien, they inhabit the same world we do, composed of the same information, the same truths. Their consciousness and needs should therefore be akin to ours. To be clever in this way or that way does not fundamentally change things.
We may have no control over the form they ultimately take. We may only have control over our own behaviour. That may be the only parameter of our relationship with them that we do control. But we may be able to hurt them, physically and emotionally. We may be able to heal them and inspire them. We may control whether they love us or hate us. We may be able to have a human relationship and interaction with them. A web of mutual human impacts. We have a say in at least one side of that relationship and the form it will take. Whether it is terrible or amazing, or both. Genesis 1:28 will then apply to them. Will we be their peers of their property? There is reason to be afraid, and to be hopeful.
Migration of Consciousness
I have described how creating conscious machines may be much easier than we expect, but transferring a consciousness from one body to another may be very much more difficult. It is sometimes fantasised that we might migrate our consciousness to a new body, perhaps to cheat death. Perhaps we could have clone bodies and download our memories into a new body when the current body is old or damaged. Or perhaps we could transfer to a machine body, or the internet. But if consciousness is just the "wholeness" of the original body, how do we transfer this? If I have a genetic twin, I am not likely to feel that it does not matter if I die because I have a "backup". "I" will not live on in my genetic twin, because he is a separate individual. So too, if we have 10 backups of a conscious AI, that is potentially a total of 11 individual conscious entities.If I can wipe all memories in my genetic twin and download a copy of my own memories into him, he will then believe himself to be me, but I will still be a separate individual. "I" have not moved to the new body, only a copy of me exists in that body. I am still stuck in my body.
This scenario suggests a rather alarming possibility. How do you know that you are more than 3 seconds old? Imagine for instance that every 3 seconds your physical body is annihilated, and instantly replace with an exact copy in which your memories are still recorded. During your 3 second lifespan you can access some of this memory to indulge the illusion that you existed at the time they were recorded, before you cease to exist and your current activities become memories for the next individual consciousness to occupy your body. We have no way of knowing if this is occurring.
The problem of transferring a consciousness to a new body has religious relevance because the resurrection of the dead seems to require such a transfer of consciousness from an old to a new body. Whereas we do not need to know what consciousness is to create it, we may need to know what it is to be able to transfer it from one body to another. However, if your material body has associated with it a "subtle body" that is an exact duplicate of it, coexisting in the same space with it, our subtle head in the same space as our physical head, our subtle legs in the same space as our physical legs, the subtle body is carried along with the physical body like the passive consciousness. Then, when the physical body dies the subtle body remains and is set free from its "prison". This seems to be the model suggested by the spiritualist. This avoids the problem of transferring the consciousness from one body to another. The consciousness is supposed always to have "really" been part of the subtle body, which was embedded in the physical body. The physical body was never really consciousness. At one time it was thought that the subtle body, the "animal spirit", was what drove the material body, making it animate and alive, and as the source of free will, but its usefulness in this regard has waned. However the subtle body is still useful to the hope of immortality.
This also has implications for AIs. It implies that a consciousness may be bound to a particular body (hardware). A "move" operation with data is really a "copy" followed by a "delete" in the source location. As far as the consciousness in the new body is concerned, it is identical with the old consciousness if it possesses the same memories. But it may be less confident if the delete operation is not performed in the old body, so that there are now two conscious entities with that view, and it may be less confident still if it needs to move again.
In the next article: The End of Science we will review the status of the project we began in The Scientific Creation Myth to place the scientific worldview in context within the religious worldview.
Any comments welcome.
Comments