Friday, 13 November 2015
SMARTS Update: Writers Beware, the Machines will Supplant Us
In the West, for millennia, intelligence was assumed to be something displayed by humans alone, a gift from God who made us in his image. Rene Descartes asserted that intelligence and the capacity to reason are properties of our immortal, ephemeral, God-given souls, nothing whatever to do with our bodies which corrupt and die. As Western science began its slow development from methodology to ideology, these notions were simply incorporated into it. For one thing, in many places in the 18th Century it wasn't safe to question them in public. Thinkers who differed had to worry about being tried for blasphemy. David Hume, the diplomat/philosopher/father of psychology who led the Scottish Enlightenment didn't buy any of it, and said so, but very carefully. His last book debunking religion had to wait until he was safely dead before it could be published and even then, his best friend, Adam Smith, wanted no part of it. It fell to Darwin, the greatest of all ground breakers, to begin a systematic investigation of intelligence in living things other than humans, everything from worms to twining plants. Yet the idea of intelligence as a property particular to humans alone refused to die. Darwin's work in this area was shoved to the back shelves of science for many, many years.
SMARTS details how modern scientists finally broke out of this religious straitjacket. Breaking free led eventually to software and chips designed by philosopher-engineers to manipulate symbols, cope with representations, and use logic as we do. It also led to tiny robots that can make decisions and solve problems without a specific program. It will lead to warfighter robots that are smarter, faster and have access to much more information than we do. First came the recognition that intelligence exists in animals other than humans, specifically our closest relatives, the Chimpanzees. Next came the understanding that intelligent behavior is exhibited by every living system, and so is as variable as the individual life forms shaped by four billion years of evolution. Regretfully, science had to set aside the lovely idea of immortal souls. Intelligence is not a God given property of humans alone, it is a phenomenon that emerges as bodies struggle to live in all kinds of difficult environments. Artificial intelligence is already emerging in the machines we build to cope with dangerous environments or to do jobs more cheaply than humans can or will.
Very few of these developments are reflected in current literature. There is a deep chasm that divides those of us who read and write stories, from those of us who practice science. To judge by the jury's award of the 2015 Giller Prize to Andre Alexis for his novel Fifteen Dogs, the literary world is just not familiar with about 80 years of scientific and philosophical exploration of intelligence. (Yes, philosophical. The mathematical logic that underlies all of our computer systems is a product of philosophical as well as mathematical imaginations.) We, who read it, think of literature as a means to travel to a zone of enlightenment: we think it carries us into the hearts and minds and experiences of others, it enables us to learn about tragedy and terror and joy and triumph without ever leaving our armchairs. One writer, John Gardner, even had a character say that only in literature can we find the answer to every problem, an example of every facet of important experience. So how can it be that this huge revolution overtaking us now, the embodiment of new forms of intelligence in machines, the expansion of intelligence and consciousness into the things we build, is so dimly reflected in literature? It is dealt with in movies. Why not in literature?
The Giller jury citation says this of Alexis' book:" What does it mean to be alive? To think, to feel, to love and to envy? André Alexis explores all of this and more in the extraordinary Fifteen Dogs, an insightful and philosophical meditation on the nature of consciousness. It’s a novel filled with balancing acts: humour juxtaposed with savagery, solitude with the desperate need to be part of a pack, perceptive prose interspersed with playful poetry. A wonderful and original piece of writing that challenges the reader to examine their own existence and recall the age old question, what’s the meaning of life?"
While I was researching SMARTS people often asked: aren't you really talking about consciousness when you say you're interested in intelligence? Aren't humans the only conscious creatures? No and no. Consciousness is awareness of self, but awareness can have many different levels from high alert to vegetative state. Consciousness exists even in brainless plants, which lose consciousness just as any animal will do when under the influence of molecules that suppress certain kinds of electro-chemical activity.
Yet Alexis, when interviewed, sometimes used the words intelligence and consciousness inter-changeably. The premise of the book includes the notion that language is something unique to humans which shapes us entirely, that the gifts of language and intelligence/consciousness that humans enjoy/endure will make dogs very unhappy. Or not. There's a bet involved between two Greek Gods. Alexis tells us that these fifteen dogs transformed by the Gods became a useful vehicle for him to explore human nature.
This is an exploration that would have been more meaningful about 100 years ago.
Alexis tells us that he became interested in dogs when, for several months, he took care of eleven belonging to a friend. He tried to inveigle himself into their pack by joining them as they howled at night, and felt he was accepted and then felt he was intruding. He obviously understands that dogs, like people, employ sounds and gestures to communicate with each other, and dogs are well able to understand basic human words, though possibly not human syntax. We know, now, that language is but one means of communication that may or may not be specific to humans and that humans are well able to communicate without it. We don't yet know if different life forms have languages of their own because we have just begun to investigate that question. For years, scientists tried instead to teach obviously intelligent animals like dolphins, chimpanzees, bonobos, and gorillas to use human languages (such as sign languages and even computer languages invented for this purpose) as if only human languages could convey meaning. When plants communicate with each other and with the animals that they use to reproduce (like us) or to protect themselves from predators, they make particular molecules to convey particular information (come hither, or hey, get off me!). Some fish use electrical signals. Octopuses use physical gestures and incredible control of various pigments in their skin to camouflage themselves, to project anger, to deter predators, even though they're color blind. Birds use tones and songs to tell competitors to take a hike. Language in other words, is just one method to convey information but there are many more, most of which we humans miss altogether (such as the voltage communication system used by tiny Amazonian fish, such as the sonar systems used for exploration and communication by dolphins and whales.)
So language or lack thereof says something but not everything about intelligence and consciousness.
One can be conscious without any sensory perceptions at all according to Christof Koch who has worked on this problem for years, along with Francis Crick (co-discoverer, with James D. Watson, of the molecular structure of DNA). And that brings me to conscious machines. If machines can reason, and machines can sense and move around in the world because we give them sensors and actuators, machines will have consciousness of some sort and may only lose it when we pull the plug or take out the battery. Some machines may have it already.
You will not find this terrain explored well in current literature. With a few exceptions (Faber, Atwood, Gibson's early work) it appears only in genres on the margins of literature -- in science fiction or fantasy. Genre stories are generally poorly imagined, badly written, with speculation to drive a plot, and wooden posts pretending to be characters. Yet it has fired the imaginations of the scientists and technologists and entrepreneurs currently creating intelligent machines, folks like the founders of Google and the designers at Apple. Since the turn of the last century, literature has mainly confined itself instead to describing the interior drives and desires of human protagonists coping with their personal experiences. You know the themes. The Marriage Plot. Love and longing. From misery to triumph and back again. The personal terrors wrought by war, evil States, parents from Hell, psychopaths. The literary world has shrunk to I and Me.
I loved 19th and early 20th Century Russian novels because their characters grappled with philosophy, geography, biology, physics, history, slavery, religions, morality, social science, war. You could read Tolstoy and Turgenev and Dostoevsky and learn what their contemporaries thought about everything. Their characters were fully imagined human beings tortured by the onslaught of both technological and ideological revolutions just as we are: these writers exhibited a passion to understand the new things coming at them even as they coped with the social strictures bursting all around them as the State they lived in unraveled. But after the Russian Revolution, literature in the Soviet Union disappeared into Samizdat, and in the West began a process of narrowing itself to simulations of human consciousness, often in harrowing circumstances. Oddly, in the same period, psychologists began to argue that the brain is too complex to study, that the mind should be treated as a black box and only behavior examined. Ethologists and comparative psychologists argued the same about animal behavior: it should be treated as a product of hardwired reflexes, not flexible like human thinking.
And maybe that's where science and literary story telling began to split, a rupture that continues to this day. Or at least it continued until DARPA got interested in how narratives might be reduced to algorithms which can be used to whomp up great propaganda in record time. DARPA funded projects which produced software now sold by companies that can write news stories without much human input. Now too scholars at the Max Planck in Leipzig and at McMaster University in Hamilton are studying the neuronal patterns of brains making art in various formats, from music to dance to storytelling, patterns which will no doubt be reproducible as algorithms useful in intelligent machines.
In other words, one day, not too far into the distant future, machines will generate stories, perhaps stories that will examine machine consciousness, machine dreams. They will most definitely write stories about us.
Writers, take note. Science is already invading our turf. If we don't write about the way the world turns soon, the machine children begat by poorly imagined science fiction will do it for us. Literature will find new Subjects whether we like it or not.