The discussionabout immortality has always been an infinite philosophical dilemma. With theprogress of technologies related to AI, perhaps we should rethink if “philosophizingis learning to die”1. Hundreds ofpeople have already opted for “cryogenic preservation” rather thansimply dying and will expect science to advance far enough to give them asecond chance to live. But if we treat death as a problem, what are the ethicalimplications of the solutions being proposed with AI revolution? At the moment,it has not yet been discovered how to attain human immortality – and it is noteven clear whether this will ever be possible. But two hypothetical optionsthat can be made possible by the AI revolution attract attention: rejuvenation technologyand mind upload.
AI REVOLUTION It seems thatquantum computers will be able to solve in seconds problems that would takebillions of years to the most powerful of today’s supercomputers. The newprocessors will allow a technological and scientific revolution difficult toconceive. But how close are we to this border? The so-calledquantum chips are the brain of a new kind of computer that seems to be able tosolve in seconds problems that would take billions of years to the mostpowerful of today’s supercomputers. The Nobellaureate and physicist Richard Feynman once stated that the laws of physics donot prevent the size of computer bits from reaching atomic dimensions, a regionwhere quantum mechanics is in control.2 The observation in the1985 paper referred to an empirical law discovered in 1965 by the engineerGordon Moore.3 Moore’s law,as it is known, states that for the same manufacturing cost, the processingpower of computers roughly doubles every year and a half. This isbecause the transistors – electronic components that represent the bits incomputers – have their size halved every year and a half.
In practicalterms, Moore’s law explains why microprocessors over the last four decades haveincreased both their ability to analyze information. A 1960s’ computer hadsomething like 10,000 transistors (or bits). Nowadays, that number hitsbillions4.
The most strikingimplication of Moore’s law is its accelerated evolution: by 2020 each bit willhave the size of a single atom. The exclamation point is almost irresistible:in the mid-1960s, a bit was the size of 10 quintillion atoms. Thistechnological law has another consequence of extreme importance. Physics thatwe use routinely and that is used to study objects with sizes of soccer balls,cars and airplanes must necessarily leave the scene. In the atomic dimension,it is necessary to appeal to quantum mechanics – this is the case of thetransistor formed by a single atom.
From this knowledge area comes terms asquantum bit and quantum computer. This changewill imply a tremendous leap for mankind. A bit of information that up to thispoint was represented by an object containing billions of atoms chemicallylinked to each other will pass into the nanoscopic domain, that is, the atomicone.
And nothing prevents a bit in the future having the subatomic orderrepresentation. At this point,one wonders: what can quantum computers do differently? The answer is: everything. Computers areessential tools for scientific and technological progress, with virtuallyunlimited applications. In fact, it is impossible to imagine society todaywithout these wonderful machines. The more computers evolve, the more theybecome indispensable. Still, thereis a kind of task that is extremely difficult – and nowadays impossible – forcurrent computers: to perfectly simulate nature itself. Scientists areinterested to simulate the behavior of natural systems, such as a chemicalreaction of a molecule in a drug, possible changes in the sea movements andatmospheric currents caused by global warming or the complex interaction amongneurons, axons and synapses of the human brain.
In thiscontext, to simulate means to reproduce in the computer exactly the naturalbehavior of the phenomenon, with as many details as possible. This is importantbecause it allows scientists to make accurate predictions, develop newmedicines and deepen knowledge about ourselves. The problem isthat if all the details are taken into account, the simulation becomes socomplex that it goes beyond the processing and storage capacity of existingcomputers – even supercomputers. Thealternative used by scientists and engineers is to simplify the problem or, asthey say in the science jargon, to make approximations. However, importantpieces of data are lost with approximations.
A quantumcomputer, however, is capable of simulating natural systems withoutapproximations. Feynman considered that nature itself is a quantum computersimulating the phenomena we observe – including us, human beings.5 REJUVENATION TECHNOLOGY With masteryover quantum technology, it would be possible to build a futuristic youthfountain by reversing the damage of aging at the cellular level. The cellbiologist and Nobel laureate Yoshinori Ohsumi, with his work in the field of autophagy,the processes by which the cell digests and recycles its own components, madecrucial discoveries that make possible the cell replacement or repair atregular intervals, what can be a landmark for definitely overcome aging. 6 In practicalterms, we may frequently visit rejuvenation clinics, where quantum machineswould perform an impossible task for humans, not only removing infected ordamaged cells without any imprecision, but also inducing healthy cells toregenerate more effectively and discarding accumulated waste in order to itssuperhuman processing capacity. This deep recomposition would turn the bodyclock back, becoming its patients physiologically younger their actual age. People,however, would remain as vulnerable as in the past to a death caused by acutetraumas, as injuries or poisoning. Rejuvenation seems to be a relatively lowrisk solution because in practice it extends and improves the body’s inherentability to regenerate itself.
However, if we are really in search of an eternallife in a biological body, we should be on our guard. Any risk of physical harmshould be avoided. MIND UPLOADOther optionwould be the mind upload – that is, scan the contents of someone’s brain andstore it on a computer.
This method assumes that consciousness resembles asoftware that operates on a kind of organic hard drive; that what makes youyourself is contained in the total amount of information stored in the brain,and therefore it would be possible to transfer that content to a differentphysical substrate or platform. With the possibility of making simulationswithout approximations, this procedure could be done safely because no originalinformation would be lost. However, thisremains a highly controversial hypothesis.
Leaving aside the question of wherelies what makes someone who he is, let’s explore the idea of reproducing thebrain in digital form. Like rejuvenation, mind upload presents difficultethical questions. It is possiblethat an upload might seem functionally identical to the brain that originatedit, but without conscious experience of the world. Its result could besomething that would look more like a zombie than a person – and even less theoriginal person.
However, itmight not be a problem at all. Another possibility would be that this would notbe a problem. As a person would be reducible to the processes and content of hisor her brain, a functionally identical copy of it – no matter what substrate itoperates on – would not be able to generate a result other than a reproductionof the original person. There areother issues. There is no way to predict what the upload itself would cause asa sensation in the mind that is being transferred.
Would this person go througha dissociation of some sort or something even more difficult to predict? And if thewhole process, which would include the existence of the person as a digitalbeing, is so qualitatively different from biological existence that the resultis a complete panic, or even a catatonia? In this case, what happens if thetransferred person can no longer communicate with others, or if he disconnects? Immortality wouldbe more of a curse than a blessing in a situation like this. Death would notseem so bad at all, but unfortunately it might not be possible. Anotherproblem stems from the possibility of reproducing a brain and allowing the copyto live in parallel with the original. If one’s uniqueness depends exactly onthe person remaining singular – which means that a fission in one’s identitywould mean death.
That is, if a person was split into person X and person Y, hewould cease to be the original person and would be dead for all purposes.However, although the original person may not survive a fission, as long aseach new version of the person maintains an uninterrupted connection with theoriginal, this could mean an ordinary survival. ETHICAL DILEMMAS Which of theseoptions would be more complicated ethically? Rejuvenation would probably be aless problematic choice. Yes, overcoming death, if this applies to the wholehuman species, would greatly exacerbate our existing problems of overpopulationand inequality, but at least we would have reasonably familiar problems toface. We can be somehow certain, for example, that rejuvenation, at leastinitially, would not be accessible to a large part of the population,reinforcing the disparity between rich and poor, and would eventually force usto make decisive choices about resource use, not the rate of population growthand so on. On the otherhand, mind upload would create a plethora of unprecedented ethical dilemmas.
Minds carried on computers could constitute a radically different sphere ofmoral agency. For example,we often consider as relevant the cognitive abilities to assign the moralstatus of an agent, even though it would be difficult to grasp the cognitivecapacities of minds that can be increased by computers that are faster and willcommunicate with each other at quantum speed7, considering that thisfact would make them incomparably smarter even than the most intelligent ofhuman beings. We would needto find fair ways to regulate the interactions between new and old domains, andwithin the new domains – that is, both between humans and mental uploads aswell as among the uploads themselves. As thetechnology related to AI increases, debates to build a framework that ensuresthat AI-enabled systems are governable; that they are open, transparent, andunderstandable; that they can work effectively with people; and that theiroperation will remain consistent with human values and aspirations must bebrought to spotlight. I believe this will one of the main roles ofinternational public policy specialists in the future: work on these key pointsto keep the mankind economically relevant in front of the AI development.
Another pointto consider is that the amazingly fast development of digital systems meansthat we may have very little time to decide how to implement these regulations,albeit minimal. What about thepersonal and practical consequences of someone’s choice about immortality? Assuming thatthis person reaches a future in which rejuvenation and mind upload technologiesare possible, his or her decision will depend on the range and type of risk he iswilling to take. Rejuvenationseems to be the most conventional option, though it may cause on people moreconcerns about protecting our fragile physical body.
Mind uploadwould make human brains destruction much harder, at least in practical terms,but it is unclear whether a person would survive (in any relevant sense of theword) if he was copied multiple times. This is a completely unknown territorywith higher risks than rejuvenation ones. Yet theprospect of being liberated from the fetters of mortality is undeniablyattractive – and it can be a way to prevail in our choices.1 Michel de Montaigne; Charles Henry Conrad Wright (1914). Selections from Montaigne, ed. with notes,by C.H.
Conrad Wright. D.C.
Heath & Co.2 Richard P. Feynman. (1985) QuantumMechanical Computers. Available at: http://www.quantum-dynamic.eu/doc/feynman85_qmc_optics_letters.
pdf(Accessed January 29, 2018)3 Gordon Moore. (1965). Cramming more components onto integratedcircuits. Electronics Magazine. Available at: http://www.cs.utexas.edu/~fussell/courses/cs352h/papers/moore.
pdf(Accessed January 29, 2018)4 Ali JaveyEt Al. (2016). MoS2 transistors with 1-nanometer gatelengths.
Science. Available at: http://science.sciencemag.
org/content/354/6308/99(Accessed January 29,2018)5 Richard P. Feynman. (1982). Simulating Physicswith Computers. International Journal of Theoretical Physics. Available at:https://people.
pdf(Accessed January 29, 2018)6 Nils-Göran Larsson; MariaG. Masucci. (2016). Scientific Background: Discoveries of Mechanisms for Autophagy.Available at: https://www.nobelprize.org/nobel_prizes/medicine/laureates/2016/advanced-medicineprize2016.
pdf(Accessed January 29, 2018)7 Ming-Liang Hu Et Al. (2017). Quantumcoherence and quantum correlations. Quantum Physics.
Available at: https://arxiv.org/pdf/1703.01852.pdf(Accessed January 29, 2018)