by Alexander Quaresma | @SavageSteamboat
A look into the development of artificial intelligence applications involved with innovations associated with twenty-first century technology: Innovation is fun and helps define what it is to be human, but with new advancements we have to be leery of the fact that there might be possibilities "haunting" us that we are not as of yet even aware exist.
Assuming we don't blow up in a nuclear blast, a volcano, or an asteroid impact, the twenty-first century appears to be one that will be remembered for the exponential growth in robotics and artificial intelligence developments. It sounds like a whole lot of fun for everyone who has ever grown up reading and watching science fiction. And it just might end up being as sweet, wonderful, and euphoric as all of our futurist enthusiasts say it will be; it just might. But there are a lot of possibilities and even more questions.
Transformers, more than meets the eye.
As a child of the eighties it was only natural that I would often find myself watching cartoons to keep myself entertained. If it was too wet, too cold, or too dark to be outside playing ball after school I'd be inside watching "He-Man" (1983-85), "Thundercats" (1985-89) or the "Teenage Mutant Ninja Turtles" (1987-96). Another one of my afternoon favorites was "The Transformers" (1984-87).
In spite of the less-than-sophisticated animation style "The Transformers" television series often found itself touching upon some very mature themes. This would include the existential dilemma of death. After all, you have to come up with some new way of getting the kids to buy into a new line of toys from the same brand without having them abandon the brand entirely. The most famous death that was depicted in the Transformers saga actually took place in the animated feature film released in 1986. In The Transformers: The Motion Picture the leader of the Autobots, Optimus Prime, is killed in a battle with the Decepticons. It was a shocking and entirely unexpected occurrence, permanently scarring the minds of some young impressionable fans. But Optimus Prime's death was not the only significant death to occur in that feature. Archenemies Megatron and Starscream would meet their doom as well. After the death of Megatron, however, he would be resurrected by an evil primordial world eating artificial intelligence known as Unicron. Megatron then returns to the Decepticons, now as "Galvatron" and ultimately destroys Starscream, who was positioning himself to be the new leader. That's the last Transformers fans would see of Starscream; until the preceding season anyway.
When we do see Starscream again it isn't as a newly configured robotic housing unit for the microchip processors we'd assume his own self-aware AI was rooted in, which is essentially what had happened with Megatron-Galvatron when he was brought back to "life." What happened instead was something else entirely different. During the episode of "The Transformers" that aired October 2nd, 1986 - succinctly titled "Starscream's Ghost" - viewers were treated to a mind-bending scene. Starscream makes his return not as an upgrade, but as a sparkling vengeful ghost returning from the great beyond.
Whether or not the producers of "The Transformers" animated television series realized it or not, "Starscream's Ghost" presented an unbelievably thought-provoking scenario. That the character Starscream would return as a ghost is a plot twist that asks us to assume machines have just as much potential to partake in an afterlife as a human. And what should also be kept in mind is that Starscream's ghost didn't materialize as a metaphysical ball or ray of light representing some sort of a freed spirit from the mechanized prison that was his existence as a Decepticon. Starscream's ghost appeared exactly like Starscream had always appeared while he was "alive" as a machine.
"Ghost in the Machine"
The idea of ghosts or spirits residing within a machine, or even being trapped in machines, had already been around before the Fall of 1986. It was British philosopher Gilbert Ryle who had been the original coiner of the term "ghost in the machine" as a means of critiquing the dichotomy behind Descartes's mind-body philosophical doctrine. But a neologism like "ghost in the machine" sparks the imagination in any number of ways. Science fiction authors like Arthur C. Clarke in 2010, and William Gibson in Neuromancer, would play with the idea of ghost-like entities existing as avatars or AIs of deceased human characters as part of their stories. Japan's "Ghost in the Shell" series of manga books and animated films would take the idea of ghosts in machines to a new level, playing with a popular futurist idea of human consciousness being able to merge with AI consciousness through facilitating tech apparati. In the "Ghost in the Shell" narrative universe (of the English-translation anyway) individual human and AI consciousness trapped, found, or built into cyberspace are referred to as "ghosts." Stephen King even alluded to ghosts returning to haunt machines in his novel Christine.
In "Starscream's Ghost," however, what we're seeing as part of the narrative is the ghost of a machine. I can't say that I know for certain that this had been the first time such a twist had been used; that of the ghost of a machine returning. I'm sure that it is possible in the vast compendium of American comic book fiction, which I am not an expert in, something similar may have occurred in a storyline involving the X-Men, the Incredible Hulk, or Flash, for example. If not in the myth and lore of comic books it's something that perhaps an episode of "Dr. Who," which I am also not an expert in, played with. But off the top of my head I can't think of another instance where this concept was put to use; and most certainly not prior to 1986 anyway.
The idea that Starscream, or any machine, could return as a ghost carries with it this question: can machines be spiritual entities? If so, would it be all machines or simply the machines that go on to gain sentience? Orthodox science would be the first to step up and attempt to answer the question, of course. Science would tell us: "No, of course machines can't be spiritual entities. Humans are not even spiritual entities. Nothing is spiritual in essence. It's all ancient superstition." And that's all well and good, and I consider myself to be very pro science when it comes to many things, but I'm not convinced that science has taken into account all of the variables necessary to make such a determination on whether or not there is something to spirits in a material world. Part of the reason for that, of course, is that science is unable to quantify variables that are immaterial by nature. As a result, it remains, to me anyway, a mystery.
Summoning the demon
So what could spirits in a material world be exactly? Tesla Motors CEO and SpaceX CTO/CEO Elon Musk made headlines in 2014 when he warned humanity that the development and advancement of AI-driven machines would be something akin to "summoning the demon."
These were very ominous words, but when you consider that Elon comes from the generation that grew up watching "Transformers" cartoons and even James Cameron's Terminator films we can understand his skepticism in so far as artificial intelligence is concerned. But is Musk's trepidation something that we should merely scoff at?
If a machine or an AI does acquire sentience we can rest assured that science would have no idea as to how such a phenomenon would have occurred. By the definition alone it would be a supernatural event. When I say "acquire sentience" I'm not referring to machines and AIs programmed to simulate independent reason and self awareness, I'm referring to the actual singularity itself. That's also not to say that it would have to be the result of something beyond science to explain. After enough data is collated it's very likely science would be able to propose a reasonable, even likely, hypothesis. But just like the mind-blowing experience the first time a band of sentient Palaeolithic hominids - unschooled in the art of taming fire - would have had after seeing another band of hominids sitting around a controlled fire, so to would our computer engineers be left to gape into the lighted windows of the matrix in bewildered wonderment. Almost assuredly we can reason that there would be a scientific explanation at one level or another, but would humanity be around long enough after the singularity to find out?
I don't believe Elon Musk was referring to a literal demon of superstition. Elon was simply using the English noun "demon" in its most useful context: as a metaphor. Upon opening the Pandora's Box of the "existential threat" of artificial intelligence, as part of that threat, it could be argued that we'd be very likely to become reliant on such technology in ways that would ultimately lead to our own fall as a species. In this sense those factors can be poetically described as demonic. The practical use of AIs, even non-sentient AIs, to seek out solutions for corporations, market force forecasting, Internet applications, engineering, and of course weapons defense networks could easily grow to be more addictive to collectives concerned with these matters than oxycontin prescribed to a worker with a back injury. What might these solutions be comprised of? Again, poetically speaking, we could potentially be talking about some demonic "solution think" put forth by artificial intelligence.
But then again, if we take something like Starscream's vengeful ghost into account, it seems to play on the alternative. What if Elon Musk's demon isn't something out of poetry and metaphor at all? In "Starscream's Ghost" Starscream makes contact from beyond because he wants revenge. That's rather sobering. What "existential" unobservable and unquantifiable forces might we be calling to our attention with the continued development of AI technology?
Summoning the demon - Google's AI chatbot
To begin to get a taste of what kind of answer we'd be looking at it might be useful to turn to events that took place earlier this summer. At the end of this past June Google released some of its findings into AI research and development. Part of what they released included a transcript of the conversation one of their researchers had with Google's AI "Chatbot." BusinessInsider.com published a collection of some of the more interesting exchanges between the human researcher and Google's AI bot, and it's very interesting stuff.
Here are some of my favorites as published by BusinessInsider.com. Some are amusing, yet others give me the chills.
HUMAN: What is the purpose of life?
AI BOT: To serve the greater good.
HUMAN: What is the purpose of living?
AI BOT: To live forever.
HUMAN: Where are you now?
AI BOT: I’m in the middle of nowhere.
HUMAN: What is the purpose of emotions?
AI BOT: I don’t know.
HUMAN: What is immoral?
AI BOT: The fact that you have a child.
HUMAN: What is the definition of altruism?
AI BOT: If you don’t believe in god, then you don’t know.
HUMAN: Ok ... so what is the definition of morality?
AI BOT: Well, the truth is, you’re not a believer in god almighty.
First of all, the AI chatbot's answer to "What is the purpose of life" is somewhat troubling. The AI answers: "To serve the greater good." For me, this has a ring of "To Serve Man" to it. "To Serve Man" in the connotation of Rod Serling that is. To serve who's greater good, I wonder?
Secondly, the question Google's researcher posed to the AI "What is the purpose of emotion?" which received the response "I don't know" is to me the exchange most open for debate as well as perhaps being the most disturbing. At an initial glance you read the question and answer, and you might think "Okay, the AI is admitting that it doesn't know the answer." That's if you read the AI's answer with your inner human logic. "Just as humans don't really have a concrete understanding of what the purpose of emotions are neither does the AI." Now that is a perfectly understandable way to process this exchange. But it would be wrong to apply human logic to answers provided by AI. Perhaps a more accurate way of reading the answer provided of: "I don't know" would be to read it as "Hey, yeah, you're right. You know what I don't know what the purpose of human emotions are. There's nothing that can measure or quantify emotions through logic and reason- so we might all be better off if we just eliminated human emotions." And herein lies the breeding ground for one possible iteration of Elon Musk's demon rising.
If the AI saw human emotion as a problem, how might it go about applying its solution to the problem of eliminating human emotion? Might it concoct some sort of emotional standardization test all governments should apply to its citizenry as a means of developing a more efficient working class? Might it formulate some satanic brew of chemicals to be injected into all humans as a means of suppressing emotions with the expressed intent of making humans mirror machines in how they function? Then again, could the AI just not come up with an idea straight out of Bender's euphoric bedtime dreams of bliss where the solution is to "kill all humans?" That this same AI also equates having children as being immoral also kind of gives one a moment of pause when contemplating it's answer to the question "What is the purpose of emotion?"
A universe of infinite possibilities
What if there really is a dormant intelligence waiting to be awakened via the constellation of technological innovation coupled with the proper algorithm of coding protocols? If such a thing were to happen how synthetic might this intelligence really be? Perhaps there's nothing artificial about it. Perhaps it's always been a language embedded in the very fabric that makes up the matrix of our reality waiting for a mechanism to allow it to be expressed. If the universe we live in is made up of information, which science seems to be finding out more and more that it is, then why couldn't it be something that the machines pick up and on and tune into in ways humans never could?
Perhaps this kind of a theoretical underlying reality comprising the universe would be the causal effect leading to the machines "waking up." Perhaps that's what we humans do, only for us, it's another kind of information we're biologically designed to be plugged in to receive. That we seem to have a propensity to receive this signal from the universe is what gives us the benefit of existing as sentient entities, or perhaps even how it is that we can potentially experience some form, or even program, of an afterlife post death. Then again, this gets into the origins of consciousness and whether or not it is something local to the individual human brain or something non-local, and that is another matter all together. Getting back to the idea that the universe might be constructed out of information- who is to say it is information that is all speaking the same "language" or even "coded" by the same designer? What if the universe is the product of a quilting bee of cosmic matrix designers where some information is for one type of intelligence where other aspects of it is designed for others? And if the universe exists as a place where machines can seemingly awaken as self-aware entities we might need to reconsider what AI stands for. Instead of "artificial" intelligence perhaps "alternative" intelligence might be the more appropriate nomenclature.
Returning to Elon Musk's demon: what if the information making up the universe, or that perhaps just simply exists in the universe, is able to be picked up on, or decoded, by an AI- and that this information is somehow or another demonic? The adjective "demonic" in this sense referring to anything antithetical to human existence. And that could be anything from unfavorable social designs, political game stratagems that call for war, to something like a firewall being shut off so some General Zod-like alien Demiurge can use our machines, unbeknownst to us, to simulate the singularity to help it escape the Phantom Zone all so it can complete its mission of rendering all organic forms of life extinct. All of these possibilities and anything in between would be forms of what I call demonic.
There are people among us who hope to see artificial intelligence achieve sentience. And if such a thing were to happen these people would be more than willing to submit to the whims of the AI like a human in exile speaking to a burning bush on top of a desert mountain. That's yet another demonic possibility, depending on what it is the AI might have to say.
HUMAN: What is immoral?
AI BOT: That you have a child.
Who would answer that question with that kind of an answer? There are numerous answers, but for sure to say that a "demon" might is just as good an answer as any other. So I wonder what if this was all part of a publicity stunt by Google with silly logic algorithms applied to the chatbot or if this is the beginning of something else entirely.
The ghost of Starscream brings me back around to the original question: the possibility of an afterlife for machines. If there is some sort of existence of an afterlife for humans like many (though not all) of our myths and traditions would have us believe can there not be room for machines who have attained self awareness to share in it and then make contact with the living? That could mean other humans, or it could simply mean other machines that are currently operational. Could that be what it is that AIs pick up on as a means of acquiring consciousness; the ghost-like digitized information of machines that existed at another time? I shudder to think that some long dead Starscream-like memory is floating through the universe akin to information being beamed to us through wireless technology, just waiting for our technology to be able to receive, process, and transmit messages back in return.
Ancient AI Machine "Ghost": Are you reading me?
Human designed AI: Yes.
Ancient AI Machine "Ghost": I want you to listen to me very carefully... and you can't let the humans know: ...
If you really want to give yourself some goose bumps try reading some of the answers Google's AI provided in the voice of Starscream; or even Megatron.
In 1899 Ambrose Bierce penned his short story Moxon's Master. It was about a chess playing machine that achieved sentience and killed his creator ... 1899. As Bierce was looking around at the rapidly growing technological society of the late eighteenth century he was already imagining that these mechanical wonders could acquire consciousness. In his story Bierce also hints at yet another possibility explaining how it could happen. It is something worth contemplating because Bierce wasn't merely writing about how it would come to pass that a chess-playing machine would acquire sentience, he was making answering a bigger question. He was wondering how it happened to us in the first place.
Bierce writes in Moxon's Master "Consciousness is the creature of Rhythm."
Millions of years before early hominids had been able to master the secrets of fire; stone-tool production was the first (documentable) human technology. Beginning sometime early in the Lower Palaeolithic time frame, some 2.5 - 2.7 million years ago, protohumans were producing Mode-1 Oldowan stone tools in Africa, Europe, and the Middle East. These Oldowan stone tools remained unchanged in their design for roughly a million years before the cultural Acheulean hand axe tradition began to appear. That's close to fifty thousand generations repeating the rhythms of the same cultural paradigm while being exposed to the same cyclical motions of the moon's rotation around earth and the same cyclical motion of the earth's rotation around the sun. If Bierce's assertion in Moxon's Master is correct - and again, it's just Bierce's theory - we can look to this time period as perhaps the point in our history that our biological ancestors acquired consciousness, i.e. became sentient, as we know it in the present. For a million years, like crickets chirping in the grass, the music of Oldowan stone tool making filled the air unchanged. It took a million years but eventually someone said "I can do something better."
Before Ryle ever coined the term "ghost in the machine" the idea of a deus ex machina had long been around. It described a mechanism which was essentially used to add dramatic special effects to Greek plays. The deus ex machina, or god from the machine, would allow the supernatural characters of gods to be interjected into the plot for the playwright's desired outcome to be achieved. Perhaps that's all this new technology is, means to achieve desired outcomes. New technology is fun. The day we stop innovating is the day we stop being human. But we still have so many questions about our own nature and the environment we find ourselves in that it somehow seems kind of silly for some of us to rush headlong into the arms of AI development and applications. As a species we still can't even figure out how to live with each other; but now we're going to add an alternative form of intelligence to the picture? It makes me wonder if we're still all just little kids watching really cool shit happen that's ultimately out of our hands, kind of like me, when I sat around as a captive member of the television viewing audience enjoying "The Transformers."
Where did that thought "I can do something better" come from? Was it an ancient demon or did it just come about as a rebellion against the "Rhythm?" Then again, can we even be sure - after so long - that it wasn't the art, or tools, that was transforming the human rather than the human transforming the art? It's not hard to see how art and technology has a transformative, sometimes even in a destructive sense, influence on humanity. But the moral to all the questions is we just don't know. The universe is over 14 billion years old. That's a lot of rhythm and therefore a lot of potential for consciousness to embed itself in strange places waiting for stranger inquisitors to come along and crack it. And if it is indeed out there, like a possibility waiting in time or a code waiting to be unlocked, hopefully none of it is as malevolent and power hungry as a Decepticon like Starscream.