Thething that I find interesting about this book is its philosophical message; theprospect of the advent of a superintelligence that is capable of our extinctionor subjugation. Most interestingly it raises the awareness as to why and howshould we control the situation before it is too late. At first the premisemight come across as a banal and hackneyed idea which is not different fromsome cheap sci-fi movie, however, Nick Bostrom tries to convince you otherwise.It is interesting to see how Nick Bostrom challenges the palpable cultural andreligious obsession with us that we, the human beings, deem ourselves as theoverlords of this Planet Earth and would stay like this forever.
We tellourselves that we are the center of the creation, but the truth is far contraryto this naïve notion. Despiteits flaws, however, the idea of Superintelligence by Nick Bostrom will appealstrongly to certain techno-futurist Silicon Valley types that I am. Here Iwould interject though. I find the example of a sudden spike in GDP over lastfew centuries, given in the book to substantiate the claim that humans canperform such great feat in achieving superintelligence as well, rather strange.But the idea that we and monkeys shared the same ancestor some certaingenerations back are cogent enough for me to believe that we are ushering intoa new era of an intelligence explosion. Nevertheless, whether you believe ornot, given the extraordinary accomplishments of humanity in the past and graveimplications it might have on us in some unforeseeable future, it warrants thata possibility of such is worth investigating. Anotherinteresting point raised in the book is that of anthropomorphizing. NickBostrom wants us to refrain us from anthropomorphizing when it comes tosuperintelligence.
It is commonplace that we attribute consciousness toanimals, trees, rivers, storms, ghosts, angels and gods. So, my answer to thisquestion why humans have hair triggered tendency to attribute consciousness toeverything around us and why we have this yearn for consciousness is simple. Ibelieve this to be a fact coming from evolutionary biology that It’s aconsequence of our hyper-social nature. Evolutionary has escalated our tendencyto model and judge others and everything and now we’re supremely synced up witheach other’s mental states. No doubt it gives us our evolutionary advantage.The inevitable side effect is the detection of false positives or falling forideas and myths as fallible as ghosts.
So, we should be cautious lest we fallfor unjustified expectations about the growth trajectory of a seed AI and aboutthe motivations, and capabilities of a mature superintelligence. Ibelieve before one digresses running different scenarios of the worldovertaking, it is only reasonable to list all the functionalities andsuperpowers one might deem necessary for such task. Nick Bostrom does the samein his book. But I feel that the idea of such superintelligence is not exploredin detail and such a short list of strategically relevant tasks is not enoughto carry out world domination.
Take for instance some of the skills likerhetoric persuasion and forecasting. They are not only quite impossible toachieve without a great deal of human intervention only via intelligenceamplification, but they are subject to chance also. In short, we areeffectively unable to predict how and when we run into such future. Nowassuming a mature superintelligence has come into being, how we are going tocontrol it. In my opinion, it fails to answer this question as well. Moreover,I find the idea of fashioning machines after human values rather vague.
Why andhow should I hope that a superintelligence more intelligent than humans inevery cognitive aspect can be enslaved by humans? Nowcomes the question of superintelligence prospect within rational bounds: is ita philosophical hoax or the greatest prediction? Today’s media, let alone theacademic circles, is bombarded by news about AI. AI seems to progress with breakingneck speed. In every circle, may it be of technology-oriented entrepreneur’selites or libertarians, most are convinced of the eventual victory of themachine over humans, the birth of a new world where human intellectual prowesshas been replaced by superintelligence.
Quite interestingly, although mediamight have convinced me in the past that a super-intelligence separate fromhuman beings is reachable within a few years, yet now I believe to thecontrary. Istrongly feel Nick line of argument lacking a rational basis. Nick Bostromstrongly believe that the human brain isn’t some uncrackable sorcery; nor arethe cognitive abilities our brains possess.
They are merely poorly understoodand mimicked. If there’s nothing magical about our brains or essential aboutthe substrate that makes them up, then we can imagine eventually buildingmachines that at least possess all the same cognitive abilities as we do. Ibelieve despite the recent advances in the field of artificial intelligence, itis still unclear how we might achieve this feat, how many pieces of the puzzleare still missing, and what the consequences might be when we do. So, we shouldrefrain from jumping on big conclusions like that. Moreover,even if I believe for the sake of argument that the making of suchsuperintelligence is ultimate eventuality, it is still in its infancy.
Numerousartists who transcend their field — from Ludwig Van Beethoven to John Lennon —have admittedly made this fact clear that they have never been able to predictwhat, when or how creations will emerge next; and indeed, or know the source ofit. Moreover, I believe that the act of bringing those “artworks” into theworld, usually ascribed to be driven by innate passion, motivation, empathy tobreak the status quo, is itself a uniquely human talent, and still not possibleto be incorporated in any of available machines. For this reason, I amconvinced that no machine, no matter how much endowed in a computational sense,will ever be able to meaningfully innovate an artistic breakthrough like JazzMusic; or a commercial one like Microsoft from scratch.
Breakthrough creativityis fundamentally organic, not algorithmic. Whilst computers and the businessesthat run on are breakthroughs; they themselves will never make them. Although Ifeel much of the idea of arriving and dealing with such a superintelligence inany point in future is debatable and futile, yet it has far more potential tohave implications in a broad spectrum of disciplines and fields. That is why Ibelieve whether such prospect is realistic or not, the human pursuit for suchmust be endorsed anyways. Inthe end on a personal note, I would like to share a muse that I had whilereading about the prospect of superintelligence. I still do not understand as towhy we revere human intelligence as a singularity; it is being called bestcreation or spitting image of God in most religious circles and now beingperpetrated as holy grail or last invention of science by human beings in thescientific world.
Instead, I believe that human intelligence and AI have theirown limitation and disadvantages, and any realistic prospect of the “seed” ofsuch super-intelligence might not grow separately out of some manmade machinebut rather a fusion of AI and humans. A cyborg generation better than simplehumans; effectively the same species yet as different as a modern man of todayfrom a caveman.