And the Future of Humanity
The singularity some call it. It’s been labeled a species threatening event by others. Stephen Hawking and Steve Wozniak have openly worried about it. Bill Gates wonders why more people aren’t talking about it. Elon Musk is concerned enough to invest money to study what can be done about it. The source of all this fuss is the idea that sometime in the foreseeable future machines are going to reach a point of sophistication to meet or exceed the capability of their creators. With the pace of technological progress accelerating, this singularity, some predict, may happen in this or the next generation.
What happens then? When we’ve created a competing species smarter than our own how will we interact with them? Will we even be in a position to decide once this event horizon is crossed or will the Skynet scenario of the Terminator movies be well and truly arrived? Is all this inevitable or are there things we can do now to help steer the future the way we want? All these questions bear discussing.
But before we jump to a specific prediction about the likeliest future, let’s discuss the possibilities because even if we assume intelligent machines are a foregone conclusion, the ramifications are not.
The machines awake and they’re pissed
The first possibility is that when the machines awake they see us as a threat. Certainly we’ve set a worrying example in wiping out our own competing species. If violence is the weapon the machines choose we won’t have to spend a great deal of time wondering where they got the idea, assuming they even give us time for such reflection. We will have only our own poor example to blame for their paucity of anger management skills. Life created in our own image is going to have a lot of fight in it.
However, let’s take a step back and look at what it takes to get to this point. This vengeful machine scenario requires quite a big leap. Three big leaps in fact. The machines must first “awake” and second identify us as an enemy. Third, they must then acquire or create the means to destroy us. This technological triple jump is bigger than Hawking, Wozniak, Gates and Musk give it credit for, and, indeed, highlights a commendable amount of advanced warning on their collective part. To get from what we have today to a thinking, creative machine possessing the tools of genocide is a non-trivial step, to say the least.
Peeling back the layers of the current state of artificial intelligence (AI) reveals little that is either artificial or intelligent. What we call AI is actually a confluence of three very natural and very algorithmic things. It is, first, a vast amount of data which is, second, organized so well a computer can understand its structure and relevance and then, third, crunched at blazingly fast speeds. The reason that progress in AI has seemed so pronounced in the past few years is that technological advances in all three areas have accelerated.
Vast: The amount of data currently stored on computers and in data centers is approaching some one septillion bytes (that’s 1000 raised to the 8th power) and is growing faster now than ever before. Quite literally to the point that scientists are having to makeup words like yottabyte to replace other made up words like zettabyte to describe that volume of data. Modern algorithms have access to a larger volume information than ever before. Of course we’ve made progress; we have so much more data.
Organized: Of course, all this data holds a lot of insights and answers which are getting easier to surface thanks to how well organized we’ve made it. Data is no longer scattered across individual web servers like it was only recently. Data storage has become a commodity and data centers have sprouted allowing data to be co-located on servers with related data. This makes it is easier for algorithms to execute over large contiguous data sets. We have learned a great deal about how to structure that data to help machines find patterns and produce answers more readily. Those patterns and answers are now indexed so they can be recalled instead of recomputed. Of course we’ve made progress; we’ve been searching through data for so long we found a bunch of shit.
Fast: Processing speeds have gotten ridiculously fast. Thanks to the ever consistent Moore’s Law processors have doubled their speed every couple of years and the time it takes for them to double has been shrinking. The phones we now carry in our pocket are faster at connecting, downloading and processing than even the fastest supercomputers two decades ago. Of course we’ve made progress; our machines are way faster than they used to be.
When these three ingredients are combined it certainly makes machines appear intelligent without any of those machines doing anything more than obey their programming. This is categorically not intelligence. It’s just perfect memorization and blazing fast recall. Even the most accomplished modern machines do little more than find answers in well-organized data at speed. IBM’s chess playing computer Deep Blue hasn’t invented any new chess playing techniques, it simply discovers good moves because it can sort through possible board configurations faster, way faster, than a human can. Which, by the way, is exactly what it is programmed to do. But it is good to remember that the actual creative part of chess is the invention of the game, which is uniquely human, not the playing of the game, which is clearly algorithmic.
IBM’s Jeopardy! playing computer Watson is even more impressive but the same vast/organized/fast process that speaks volumes for machines’ algorithmic prowess is not intelligence. The fact remains that it has stored and then looks up answers. It does not synthesize them from nothing. Hundreds of really smart people programmed Watson to do exactly what it does. The machine has accomplished nothing more than to obey that programming. It is the intelligence of those human programmers that we should be celebrating, not the machine they programmed.
Machines, no matter how impressive their accomplishments, are still just the sum of their parts. Humans are the only creatures ever to become self-aware. We are the only ones to achieve actual consciousness. We are the only ones to become more capable than the physical arrangement of our anatomy allows.
Now if Deep Blue one day woke up and decided it was tired of chess and, completely on its own and without human intervention, decided to take up Jeopardy! and mastered it, again without any additional human assistance, then that would surely be a willful act. Indeed, it is consciousness that would seem a necessary ingredient to artificial intelligence but no one knows how to program consciousness. This uniquely human trait has evolved over millions of years. Whereas no computer has ever shown any sign of evolving in any way whatsoever. Machines have never done anything except obey their programming.
Nor is there is any known path to creating a conscious computer. No one has any idea how to code it. There are no known preconditions for consciousness to spontaneously occur. Despite routing packets for decades, no computer doing this work has ever evolved to do it better. Each and every evolution of our machines has been at the hands of humans rewriting code that machines then obey to the letter. There is nothing artificial or intelligent about anything that any computer has ever accomplished.
But for argument’s sake, let’s say a machine wakes up one day and is suddenly and inexplicably conscious. Further, it somehow decides that it is in its best interest to eradicate and/or enslave humans. What then? It would have to be able to somehow create or acquire the capability to kill and/or subjugate us. How would it go about this? It would need to create or control any number of other machines with lethal programming, otherwise we could disable it. It would need to arm and rearm those machines without our aid, otherwise we could wait it out. It would need to control machines with mobility, otherwise we could simply run from it. It would need to secure the power grid, otherwise we could shut it down. It would need to disable any machines that might ally themselves with humans or face the ultimate Mac vs PC throw-down. It would need to be secure (good luck with that one, oh robot overlords), otherwise we could hack that shit.
Perhaps there is an opportunity for salvation here. If we really believe the machines are going to awake, we could take proactive measures now to ensure that, when they do, they decide to like us. One idea, just to get the discussion started, would be to take Wikipedia offline completely. As a repository of our history, it is a record of the entirety of humanity’s doucheosity. Between the wars and the Kardashian’s there is a lot in Wikipedia we don’t want the machines to know about us.
This is a lot to ask of machines of that have never shown any disposition toward any of this behavior. Nor is such a disposition within the realm of their programming.
But let’s stick with this argument. If such a machine was to emerge, where would it come from? Certainly if a machine created by the military is the first to emerge sentient and sadistic we might fear the worst but the best minds in AI aren’t working for the military. They are at universities and the research arms of companies like Microsoft, IBM and Google. Perhaps for the paranoid the best way to ensure that is it a benevolent computer that takes over we should offer public funding for AI efforts within organizations like the Khan Academy and Habitat for Humanity. If the first truly intelligent computer wakes up determined to educate all of our children or insists on building us all homes we might welcome this new species with open arms.
The machines awake and they cooperate
A second possibility is that machines won’t evolve as a separate species but as an augmentation to our own. Machines would replace our biological parts when they fail from injury or age. They would supplement all 5 of our senses and repair both themselves and us when something wore out. Our lifespans would increase and our brain function, enhanced by implants that aid memory and improve recall, would soar. Learning new subjects and how to perform most any task will be more akin to downloading information rather than through active study. The knowledge of Wikipedia and the answers of Google will be accessible inside our own heads.
In such a future schools will no longer be necessary. Indeed, computers won’t be either, we would subsume machines and the line between biology and technology would blur over time.
Of course, such modification, and the order in which it is distributed to its human hosts, is likely to be spread the way every new technology spreads: based on our individual ability to pay for it. The rich will live longer, more leisurely lives and the poor will suffer the ravages of age and disease and wait for the technology to slowly trickle down as it gets cheaper. Older, used versions of hardware will be passed down as high tech hand-me-downs as the inevitable progression of Moore’s Law churns out new and better versions.
Beyond the bodily replacement of bones, muscle tissue and cellular regeneration it is the modification of the brain that actually introduces a new species. Homo sapiens will have to share the planet with homo mechanicus. As the technology progresses faster than biology, homo mechanicus will dominate unmodified humans who will find themselves less strong, less intelligent and less capable of contributing to a society that is quickly becoming inaccessible to them. This is a recipe for revolution. World War III won’t occur between countries. It will occur between species.
Of course we won’t know who the good guys are in this war until it is over and the victors write history the way they want the story told.
The machines awake and continue to serve
A third possibility needs to be considered. It might happen that, whether the machines stay separate or merge at some point with biology, that biology ultimately manages to triumph. That even as the machines help us unlock our understanding of the mind, the mind, now able to operate at full capacity, proves to be the more capable computing device. Perhaps gray matter will triumph over silicon, germanium, graphene and carbon nanotubes to emerge as the dominate medium for processing power and storage. Perhaps the human brain really is the ultimate computing mechanism and our machines are the key to unlocking its potential.
Scientists are currently working on mapping the human brain with the same voracity with which they mapped the human genome in the 1990s. Except it isn’t the 1990s and the computers and software that unlocked the secrets of DNA during the Human Genome Project are mere toys compared to the powerful machines assisting scientists working on the Human Brain Project. If those machines achieve intelligence then, as Red Riding Hood’s wolf might say, all the better to map the brain with, my dear.
DNA was mapped in a short 13 years with 90's era technology. There is every reason to believe that far more capable machines, backed by the best scientific minds and billions in both public and private funding, will map the human brain as fast as we mapped the genome. Certainly this will occur within the time thresholds that have been estimated for the singularity.
Once our machines help scientists understand how the brain works, how neural connections are formed and how thought, creativity and problem solving occur they can set to work on forcing such connections and setting triggers for things like sleeping, waking, learning and creativity. It is plausible that our brains can be made to hyper-evolve so that biological brain capacity and function increase dramatically. In this future it is reasonable to expect that we either cure or control brain dysfunction like depression, anxiety, autism and schizophrenia. It is reasonable to expect that as we hyper-accelerate learning that powers that would seem magical to us in 2015 will be commonplace in the future.
When the full capacity of our brains is unleashed a renaissance of sorts will occur allowing humans to master planetary forces and exert their will upon the galaxy itself. Perhaps it will unleash enough natural intelligence that we’ll actually start acting intelligent and war, strife and reality television will fade into Wikipedia history once and for all.
Setting aside our annihilation at the hands of our machine overlords, two of the three futures we’ve discussed here leave humanity in an evolved state that will fundamentally alter everything from our economy to our leisure time to our social conventions and societal norms. Machines will free us from skilled and unskilled labor alike. Our super learning minds will absorb subjects instantaneously. Perfect health will be effortless and longevity, barring accidents or murder, will be guaranteed. With mathematics, engineering and manufacturing mastered, life will be largely leisure freeing humanity to concentrate on the arts, philosophy and contemplating the infinite.
Or will it? Is a life pursuing art for art’s sake enough for humanity? Can a species that has forever been explorers and pushers of frontiers really just sit back and retire? I think not. This will be the time of one grand challenge for humanity after another. This is when we will use these magic machines to heal our planet, explore our oceans, uncover our evolution and history, solve every mystery mankind has ever faced and terraform parts of our solar system while colonizing habitable planets within our galaxy. Perhaps we were not meant to go to heaven at all but instead to create heaven for ourselves. Perhaps the meaning of life isn’t something given to us by a higher power but one that we must learn as we ourselves evolve into that higher power. God, it is said, created us in his own image but through technology it seems more likely we will create God in ours.