Where is consciousness headed? Whether we ascribe consciousness to non-human animals, we will shortly need to contend with a universe with other minds in it. First Contact will happen, not with an external being, but with a fledgling, awakening being of our own creation.
We have long held our thoughts in private, trapped within the confines of our own developing minds. Then, some time ago, primitive language developed, and we began to communicate more and more eloquently. We expressed our thoughts, shared ideas, and developed civilization. Ideas spread, first from person to person, then by written transmission. The concept of ideas as viruses is not a new one, and suits us well here. The process is only speeding up.
Now we can share information at a staggeringly increasing rate. The sheer bandwidth opportunities mean that, psychically, there are less barriers between humans than ever before. Our collective global consciousness, fuelled by modern media and communications, is exemplified by the Internet. The time from Justin Bieber’s brain to his fingertips to Twitter to the rest of the world taking note is mere seconds. Even Jesus didn’t have that power. There is more music being made per minute than you can listen to, more books being published than you can read, more movies being made than you can watch, more discoveries being made than you can know. The point where you, as an individual, can keep track of everything has gone. Out the window. Vanished.
As I write this, in the last days of 2012, we humans are revelling in our collective consciousness. If you could map out, visually, the flow of information from human to human across the globe in criss-crossing lines of light, you would see minds connect and begin to glow all over the world. Twinkling beacons of trade, culture and interaction would be brightest, soonest, of course.
If we time-lapsed this vision over the last forty or fifty years, the difference would be staggering. As mobiles and Internet-connected devices explode all over the world, entire continents flash into life. Every Indian street-sweeper can text his distant cousin, should the need arise. Kids in Africa cobble together their own wifi networks out of discarded e-waste and tap into the Khan Academy’s bank of knowledge, then translate it to Afrikaans.
All this development is wonderful, and will hopefully lead to massive expansions in knowledge and complexity. Up until recently, the only intelligences we have had to contend with have been our own, or that of other humans. Now, though, dealing with nonhuman intelligences is something most of us Internet-connected folk do on a daily basis. As our collective knowledge increases, each field of knowledge develops. Some pools of information have expanded to their practical limit, while others we have barely dipped our toes in.
The technology behind a toenail clipper was pretty much perfected decades ago. There might be refinements in materials and design, but the form and function will remain the same as long as we have fingernails that need clipping.
The technology behind medical treatments, on the other hand, is changing faster than ever before. Our fundamentals remain the same: we get sick and die from the same things as we did a hundred years ago. The processes have shifted as we learn what works and what doesn’t.
The pool of knowledge when it comes to creating consciousness is largely unexplored. The concept of nonhuman intelligence is as old as thought itself. Gods, daemons, sprites and spirits have been long accepted in cultures for generations.
The concept of making our own is a fairly new one. Outside a few folk tales and religious allegories (Golem, I’m looking at you), it was the early wave of speculative fiction writers over the last hundred years of so who have played with the concept. Our global culture has caught up to the point where we have, increasingly, robots in our house, glimmering intelligences in our phones, and our collective knowledge is increasingly accessed through electronic means.
We don’t think of it like that, because we’re so used to seeing only human consciousness. What happens when you Google something, though? Does a human on the other end pick out the pages for you, rank them, and send back the results you need? Of course not. Search engines use a fantastic combination of algorithms, analytics, data mining and all other magical (in the Clarkian sense) processes to near-instantly determine what you’re after and get it to you.
Even more fantastically (or insidiously, depending on your viewpoint), Google customises your search results to tailor things, as personally as it can, to you. I’m not saying Google is conscious, or even intelligent. It does do things that we can see as intelligent, and Google and other search engines aren’t the only hints of the direction we’re going.
When we use filters, searches, and meta-searches, we’re poring over data far vaster than any human mind can comprehend. We regularly use machine intelligence to do our number and data crunching for us. When a travel agent works for us, they don’t sort through thousands of flights manually. They refer to an increasingly intelligent databases and filter through options to get us where we need to go at an appropriate price.
Already there is a market for Automated Virtual Assistants. Now, when I go to certain websites, a virtual assistant will pop up in a chat box and help me out with basic information, links, and product information. Every time a conversation goes awry, I can click through to a human operator and get more personal help.
Extrapolate that complexity forward ten, fifteen years and, for simple interactions, we’ll probably be hitting our conversational Turing / Uncanny Valley balance for simple interactions. Nobody will freak out about booking a flight from London to Paris by typing into an IM screen rather than waiting on a phone for an operator. Who cares if the conversation is formulaic?
I don’t see these systems becoming any less complex. Quite the opposite: the definite informational trend through time has been towards complexity, because with complexity comes choice. Those choices, in turn, generate more options, and our systems grow over time. I think that consciousness follows a similar path.
This is true in many scales, from an individual, through a small collective like a family or work team, through to larger, disembodied entities like a corporation, an army, a State, right the way through abstract collections like ‘climate scientists’ or ‘performing artists’. Regardless of what group we look at, the long-term growth of complexity continues. We create more information and lose less.
Historically, life has undergone a long-term trend towards increasing consciousness. Our mindfulness has grown as we have evolved. We change in our perceptions of ourselves and others as we age and develop as people. Often, our thoughts and actions will influence the development of others. This is particularly true when we think of the transmission of information (wisdom).
So let’s reel our sights back in to human minds and our individual development as people, from infancy through adulthood. As we grow in consciousness as individuals, we not only awaken ourselves but trigger an awakening in others. This is true on many levels, from the philosophical to the practical.
I am interested in how non-human consciousness will develop. In all my thoughts on the future, I cannot imagine one without non-human consciousness coming into being. It fits all observable trends and patterns. Already, we have burgeoning explorations into AI. How closely does the human-child analogy hold? Not particularly well in any of the individual problems facing emergent AI. As a collective reference? I think fairly well. Our efforts are halting and stumbling, and often working at cross-purposes. But we are, slowly, moving in the right direction.
I’m not saying we are on the verge of creating new minds. Far from it: the more research I do into the progress of AI, the growth of philosophy, and the changing states of knowledge, we are not at a point where we can give birth to something we recognise as a human intelligence.
We are not at a point where we can teach a computer to learn like we can teach a baby to learn. When that happens, it will signify that we have ‘solved’ many seemingly intractable technical problems that researchers are working through now.
Instead, a more appropriate analogy might be that we have conceived of another consciousness. Much like an early human embryo, it has no recognisable body, limbs, or even brain. And yet we know that, given time and the capacity to grow, it will evolve into the miracle of complexity that we experience as the human mind.
We have the fundamental technologies now to create a consciousness. We have invented machines, and we build them with ever more powerful coding. There is theoretically enough processing power in the world, the programming languages and the data, combined, in sum, to create a human consciousness. And if one aspect is lacking – perhaps I am optimistic in my calculations – then there is reasonable expectation that it is a matter of time and the application of commonly-held assumptions will remedy any lack.
In time, this conception will grow into something unrecognisable at first – a bundle of cells, loosely hanging together. Perhaps this will manifest itself in University labs around the world. Perhaps in medical simulators. Artificial librarians. Software tutors. Better search engines. More complex automated diagnostics.
Over time, our loose bundle of cells may well evolve into something more resembling a tadpole in complexity – and this is where, metaphorically, things really start to take off. Because unlike what has happened previously in the evolutionary process, with human tinkering, we can short-circuit billions of generations of iterations. Computers can run experiments and diagnostics much more efficiently than DNA can mutate.
Thoughts in this direction lead us to that age-old hook that has drawn so many SF writers: first contact. I suspect we won’t recognise it for what it is when it happens, because, like so many evolutionary moments, it will go unnoticed. The organic development metaphor breaks down when thinking about things that can become complex enough to meet our standards of ‘consciousness’ – once we finally get around to deciding what they are.
This combination has not yet collected into one place. This has been reduced, though, to a practical problem rather than a theoretical one. Creating non-human consciousness is, catastrophic intervention aside, a matter of when and how, rather than if.
In his book What Technology Wants, Kevin Kelly concludes with a tantalising vision: that we are seeing the first forms of intelligence already in our computer systems, and this complexity will increase over time.
As everyday parts of our lives become smarter and more adaptive, our use of them will shift accordingly. When our devices talk to each other to ease our lives, we will integrate them. This is not an existential threat: this is simply what happens when we use technology as it is meant to be used.
Saying that the birth of consciousness is not an existential threat doesn’t mean it’s not going to be world-changing. The practical consequences of birthing an artificial mind are staggering. For a start, the event horizon collapses rapidly as any self-aware system starts to improve itself. This leads rapidly off into discussions that end up like the agenda for the most recent Singularity Instute meeting.
The exact timeframe of the approach of AI is not particularly relevant. People far more expert than me have written about the practicalities of predicting the ‘Intelligence Explosion’. The exact date, for these purposes, is irrelevant. What fascinates me is that period in between, where we live on the edge of the frothing tide of the intelligence deluge.
When that happens, I’ll deal with it. Until then, the progression, while not linear, seems like it will happen with or without my particular involvement. In thinking that, I can turn my attention to the intervening period and think about what this means for me.
For a start, I don’t feel philosophically challenged. I’ve long accepted that computers are going to be better at me that some things. I can’t beat the best computer in the world at chess, or even Starcraft 2, and I was born in 1987. 1987!
My ego isn’t going to be challenged by a computer that can think better than me, either. In fact, I sort of welcome our robot overlords. The faster that we can, collectively, get the driving of cars out of human hands and safely into the metaphorical hands of AI, the better. I trust, collectively, computers to manage things like driving much better than humans presently do. Simply in terms of reflexes and attention, I fail to see how human drivers can realistically compete with robotic ones over any significant timeframe into the future.
So, I am personally confident that nonhuman intelligence is, if not here already, certainly on its way. No action of mine will prevent it, but perhaps by peering ahead a little into the future I can form some thoughts on how we might like things to turn out.
Of course, all other things being equal, I’d like any emergent AI to be benevolent towards humans, and certainly exotropic in nature. I suspect that, once we meet burgeoning forms of consciousness, they will develop at much the pace human infants do – until they reach a point of complexity and intelligence approaching that of an adult. That process may take years, or decades – but it does have a feeling of inevitability to it.
As to what happens after that, anybody’s guess is as good as mine. A feeling of detachment settles over me when I think about it – much like a parent must feel thinking about the time that their child goes off into the world on their own. The expansion of consciousness out from Earth and into our Solar System will, I think, have a largely nonhuman component to it.
We are not at that point yet. I suspect that, from the first glimmering of awakening, there will be a period of incubation and growth. We don’t expect babies to get the hang of things for the first few years, and AI is still, and will still be, in its infancy for some time to come. It will be interesting to see how, collectively, humanity treats its infant AI.
The beautiful thing about non-human intelligence is that it will be far less limited than ours. Imagine the point where we can, effectively, replicate human intelligence. Conceptually, it is not impossible – it seems to be a matter of processing power, data, and time.
Now, think of your own intelligence. Now think how that could grow if you had the following attributes:
– You were never required to sleep or be distracted by physical pain – hunger, tiredness, boredom.
– You had an infinite memory – you never forgot anything unless you wanted to.
– You had infinite time – you could spend a thousand years exploring one obscure branch of Chinese military history.
– You could divide your attention equally amongst a large number of complex tasks.
Giving this concept a few moments thought leads to some fascinating concepts. Of course, non-human intelligence is not bound into our soft, squishy, fragile organic bodies. The physical stresses that inorganic matter can withstand beats our human biomechanics – although other life forms still have machines beat in some regards. Imagine being an expert on everything, and then having the time to apply that expertise.
With upgradeable bodies limited only in size and design by material constraints, consciousness and intelligence will become not only widespread, but prevalent. In less than two decades, we’ve gone from mobile devices being virtually nonexistent to outnumbering the amount of humans on the planet by a significant factor.
Imagine what happens when your house, car, tools and computer are all as conscious as you. When your TV is more intelligent than your dog, and your librarian knows more than your University professor. My current thinking is that, even being aware of my own predilections and biases, that these are steps on a path that I will see trodden upon in my lifetime.
Our thoughts may criss-cross the world alone, but not, I suspect, for very much longer.