Love and Beauty in a Time of Machine Intelligence- Published in The European, Jan 2015
Why we find beauty where we least expect it.
Last year at a conference titled Transhumanist Visions, I presented on a question that had perplexed me for the last few years- the idea of whether a futurist technological age - some sort of singularity era - was aesthetically beautiful. As a philosophy student I had debated the definition of beauty with my classmates, whether the Kantian idea that beauty might be a symbol of morality could be correct. In the Critique of Aesthetic Judgement, Kant describes how beauty is universal- the only experience on this earth that can be felt without a need for communication. In this way beauty gives humanity a 'sensus communis' when we see something aesthetically pleasing.
That said, for me beauty has had absolutely nothing to do with harmony or morality. Instead it has always been representational to me of the feeling of 'curiosity'. Something aesthetically beautiful piques our interest, causing us to want to understand it. We look for meaning in a painting we find alluring, just as we look to hear different things in a piece of music we find sublime.
It’s really about timing.
A study from the University of Montreal found that volunteers under fMRI and PET imaging experienced releases of dopamine whilst they listened to their favorite music. This isn’t particularly surprising - we would expect the experience of beauty to link feelings of rewards and pleasure that are associated with dopamine. The more interesting finding emerged from a close study of the timing of the dopamine response. In the study it was found that the greatest dopamine activity was experienced in the seconds before the climax of the experience. Fascinatingly, we actually get the high from this anticipatory phase, set off by temporal cues signaling that a potentially pleasurable sequence is about to come, playing on our senses of desire and reward prediction.
Creatives have played on this psychological phenomenon throughout time, continuously teasing us with suspended expectation. In the early 1990s, George Loewenstein of Carnegie-Mellon described curiosity as something rather simple: It comes when we feel a gap between what we know and what we want to know. In reality, there isn’t much cognitive difference between the concepts of ‘anticipation’ or ‘curiosity’. In both cases we are looking for something, even if we might not know in advance what it is.
As I listen to Max Richter’s re-working of Vivaldi’s Four Seasons, I am teased by his musical score in the pattern-recognizer which is my brain. I try to keep hold of the overriding pattern I’m familiar with, whilst the underlying notes tease me into different directions. Music can be a form of subtle chaos. The best composers play on anticipation and curiosity, building up expected patterns and then tantalizing us with crescendos and climaxes we neither anticipate nor expect. I've listened to Richter’s Four Seasons hundreds of times, each time looking for a detail I might not have noticed before. The beauty of the piece leads me to attempt to listen closer, to try and understand the patterns.
The notion of building order out of chaos can be troublesome because almost all things in this world represent some sort of pattern to be understood better. Natural fractals — irregular, self-similar geometry — occur virtually everywhere in nature: in snowflakes and flowers, in riverbeds and within our own bodies. The awareness and admiration of complexity can be overwhelming. We realize the superficiality of our understanding- the fact that our perception only represents one specific layer. As a child I would notice and admire patterns everywhere, from leaves to bird wings, to my mother’s cooking recipes and in cracks in the concrete. Stendhal’s definition of beauty as ‘the promise of happiness’ pleases me because it fits the model of beauty as curiosity. Trying to understand patterns is an attempt to reach some sort of higher-plane - some sort of contentment.
Mathematicians describe a feeling of euphoria from creating or recognizing an elegant proof. Mathematical elegance indeed appeals to the pattern-recognition idea. The simpler and cleaner the derivation of the pattern, the more it pleases us. The mental response from the build up and climax within a piece of music isn’t that different, cognitively speaking, to the working out of a formula. Each represent our attempt to pattern-match and gain understanding of the complexity and chaos that surrounds us.
If anticipation and curiosity are representational of beauty, and vice versa, then thinking about the future, by necessity, is a beautiful thing. We anticipate and ponder curiously on a future of knowns and unknowns. We look back at history and look for patterns we might understand and build from. We continuously attempt to build models as we try to make sense of the complexity of the entire universe; of how it is now, how it was, and how it might one day be. We might imagine that the whole of life could one day be represented by a mathematically elegant formula; an abstraction so clean and superior that everything we’ve ever imagined could be represented in some sort of perfect model.
When we contemplate a variety of future scenarios, we know just enough to know that we want to know more. We anticipate the road that lies ahead with a mixture of sorrow and elation, dwelling on prospective unknowns that we will never see. The climax of the future always lies ahead, it always represents a time still to come. No-one ever reaches ‘the future’ - it’s the definition of anticipation. And for these reasons I will claim that the future does not ‘need’ beauty, it is in fact always a conceptual representation of the beautiful.
Creation is the selection of order from disorder
At University I went off-topic one semester, after reading G.D.Birkhoff’s 1933 publication Aesthetic Measures. Birkhoff presented the idea that the aesthetic measure of an object can be defined as the quotient between order and complexity. In the late 1960s, the German Philosopher Max Bense built upon Birkhoff’s idea that the process of creation involves a determined repertoire of elements (such as colors, sounds, and phonemes) that is transmitted to the final product. The creative process is selective (that is to say, to create is also to select). Creation is the selection of order from disorder.
Building on Birkhoff’s theoretical frame, Bense aimed to create rational aesthetics free from subjective speculation and grounded on a purely empirical base. For Bense, the physical world heads toward chaos (entropy), whilst the world of art heads toward order (negentropy). Things that are aesthetically beautiful are those that successfully represent to us concepts of negentropy. Philosophers contemplate the criteria for a beautiful life. The question I ask is whether a beautiful life is one that simply contributes towards the emergence of order over complexity?
Let us consider how this has been achieved historically. In the 200,000 years from which modern humans evolved from archaic humans, we’ve experienced major paradigm shifts in our manifestation of order over complexity. The creation of language turned consciousness from a slippery mess into a functional tool. With the ability to create transferrable representations of concepts and ideas, humans were able to pass information from one individual to the next, generation to generation. The timings of the origins of language are debated, but regardless of the timing, the construct of language was potentially the first singularity experienced by humanity. It created an intelligence explosion, extending life-span through shared information that educated us all. We successfully created a tool to build order out of the complexity that surrounds us.
One of the most beautiful ideas of all time
I draw parallels between the preceding era of the birth of language to our current moment in human civilization. We live in the pre-era to a second major singularity, the biggest paradigm shift in order over complexity. Through the development of Artificial General Intelligence (AGI), we will create ‘thinking’ machines that will be able to apply, build, and scale intellect across broad spectrums. We will fast forward thought in a way never seen before, allowing us to understand the foundations of disorder and complexity, and we will be armed with the ability to to manipulate out of it to best fit our wishes. Ben Goertzel has summarized the aim of building an AGI as a system that in principle could achieve nearly any goal in any environment given sufficient time and resources.
The creation of AGI will lead to an intelligence explosion, as intelligence is defined into a fixed scalable formula. An ultra-intelligent machine could then go on to design even better machines. As I.J. Good famously stated in 1965, the first intelligent machine is the last invention that man need ever make. With a true definition of intelligence, we will create the most magnificent order over complexity ever achieved in the history of humanity.
In our current paradigm, we use our individual intellect to pattern-match within our own minds. Over the internet and through our global communication infrastructure, we are now able to scale our pattern-matching to a universal standard. Sadly, we still build on superficial principles which cause our abilities to have rigid constraints, as we have not yet defined the very things which we are scaling. We do not yet truly understand what intelligence is, even though it’s our foundational tool to permeate disorder and turn it into order.
The work around a formula for universal intelligence is, in my opinion, more aesthetically pleasing or promising for human potential than any painting. And although it might seem counter-intuitive to apply beauty to the worlds of computer science, once intelligence is definable and scalable, we’ll hold in our hands the tool to pattern-match anything. If beauty is the anticipatory phase within pattern-matching - the dominance of order over complexity - then the creation of machine intelligence is undoubtedly one of the most beautiful ideas of all time. We don't know how far we could go.
The battle we face is love over entropy
The development of superintelligence will finally provide us with the ability to override every other system. Humans will have the ability to overcome the limitations that have always plagued them. Aging, disease and natural disasters could all be overcome. We will go on to learn that most of the truths we assumed about the fundamental basis of human existence were always, in fact, malleable. It’s hard to visualize the after-effects of such a paradigm shift. We don’t yet know what humans are capable of.
And so I go back to Stendhal’s definition of beauty as the promise of happiness. With superintelligence we could create the tool to potentially deliver Stendhal’s definition. We maybe hold the power to manipulate every molecule that surrounds us, to deliver humans to their next transition, abolishing every constraint that we have ever faced before.
When I look at my loved ones - and can there be anything more aesthetically pleasing than those individuals that we love - I realize the battle we face is love over entropy. Deep care for one individual human reminds us of the scale of all the potential disorder, all the global diseases, sufferings and deaths. We understand so very little, and at this time are only able to pattern-match on the superficial. The cells that make up our children, the DNA of our parents, lie on unknown paths, within unknown frameworks, unregulated. Do we not, individually and globally, want to understand things better? For the sake of future generations we should be fighting for a complete understanding of the systems that define us, of the contexts in which we reside - our bodies, our minds, our planet, our universe.
Once human-level machine intelligence is demonstrated and progressed, we will wonder why we had spent so many years not making it top-priority, whilst our loved ones died of illnesses, of cellular disorder, or natural disasters so easily understood and solved by machines able to think in ways we never could. Eventually it will be plausible that all that ever existed and could exist could be modeled and ordered. We might live in chaos, but the universe is humanity’s greatest canvas.