logo

Part 2

Even if AI will not entirely replace human composition, it looks set to have a significant impact on it. What does the term composing mean in the era of AI, do you feel?

Similar to the last question, I think there will be a beautiful symbiosis which will give birth to new forms and pathways that would have been closed to us as a species were it not for the silicon-minds we gave rise to.

In other words, that region of the possibility space only became available to us once we were able to design another form of intelligence capable of traversing that space with us.

Can you see a future where AI could make aesthetic judgements as well?

Well, I would say this has already been the case for some time now.

In addition to the most overt example of AI robotic arms that create paintings, the various algorithms that ‘suggest’ recommendations on sites like Pinterest, Amazon, Spotify, Netflix, and of course, dating apps, are indeed making aesthetic ‘decisions’. Whether we want to use the term ‘judgement’ is a much more complicated philosophical issue however.

If we look at something like OpenAI’s Dall-E - which takes a natural language description from the user as input, and outputs an image based on that description, could we say that ‘aesthetic judgements’ are not being made? Of course, one could argue that in order to render any kind of judgement, one must have ‘understanding’. But because we view ourselves as pretty much the only organism on the planet that has evolved to enjoy art, that understanding really implies ‘human understanding’.

This then by definition excludes AI from being able to render what most humans would consider to be judgements. But I find this argument quite weak when one considers just how much of nature - including much of our own cognitive processing is better described as ‘competence without comprehension’.

For an overt example just think about the computer or mobile device you’re using to read these words. Digital computers, smartphones, and all manner of devices that rely on them like MRI devices, GPS satellites utilize various aspects of quantum physics - but ask any quantum physicist and they will be the first to tell you that while we can manipulate various features of quantum physics, not even the most credentialed amongst them would say that we truly understand the quantum world.

In more subtle ways, we can look at other organisms in nature - if you’ve never seen a termite colony, google it and just look at what an architectural marvel it is. Once you realize not only does it aesthetically look like something the infamous architect Gaudi might have designed, but that its design also allows for temperature regulation against vacillating exterior temperatures, one is left wondering if there are little termite architects and engineers inside with mini blueprints.

But all of these individual termites have specific functions they carry out, and while no single termite has the overall conception of what the final structure will look like, by following simple local rules individually, they give rise to something that any thinking person would have to admit looks like ‘design’.

Is there ‘comprehension’ or ‘understanding’ there? There is clearly competence. And that’s one of the most important features of nature at all levels, from successful and elaborate courtship routines throughout the animal kingdom to interactions between the cellular machinery inside your body, nature doesn’t require ‘understanding’ in order to be competent at propagating itself.

In a more philosophical, and pessimistic sense, I would argue, we as a species have not shown that we understand much when it comes to our entangled relationship with nature. We stumble forward through time with the misconception that we are somehow separate from nature, but this is pure human exceptionalism rearing its ugly head. Yes, we are unique in many ways, but in less ways than we think, and most importantly, and this was Darwin’s great contribution to our collective understanding, that our seeming uniqueness is merely an evolutionary step - not some gift due to our inherent ‘specialness’.

Returning to your question about whether an AI system makes value based judgements of whether a certain image is more beautiful than another would depend on the set of input parameters that system was trained on. If humans are feeding in that data - which they are - and humans have had a few millenia of creating works of art and categorizing them - which they have - then there is bias toward what homo sapiens call beautiful that is literally being fed into the machine as initial input.

So, my long-winded answer is yes and yes. Yes, it is capable of doing so simply as a brute fact of having our biases fed into it, but also that algorithms like Dall-E are clearly making aesthetic judgements even if we don’t want to grant them such agency because it’s more comfortable to think of these algorithms and the machines that employ them as ‘mindless’ in every sense.

In which way can AI potentially contribute to the resolution of issues facing humanity? In which way do you also see risks, both for the arts and humanity as a whole?

I think the ways in which we’ve conceived of AI and its use in society has been so incredibly small-minded that it speaks more to our own unimaginativeness, or perhaps what we prioritize as a collective society.

Think about the sectors that really drive and utilize AI - they’re defense, finance, and consumerism. In at least two of those three, defense and consumerism, AI is vastly employed as a powerful tool for surveillance - to track potential threats, (threats of course being judged by the perceiver employing the AI) and in consumption to track potential consumers.

It’s funny that much of the way we conceive of AI follows the trajectory of the agricultural revolution - particularly our domestication of specific species. We have over time artificially selected the various breeds of cats and dogs that have free reign of our homes today, then we have species we’ve bred for livestock whether for work or for dairy or meat. And the final category includes those animals that occupy some of our most visceral fears that for a variety of reasons are beyond any sort of domestication.

Similarly, we have domesticated AI systems that we enjoy using and aid in mainly entertaining us, AI systems that are foundational to making labor across various industries more efficient and powerful, and we often speak of that final uncontrollable AI which we might accidentally unleash as our final invention, that untameable beast that would forever remain beyond our control and replace us at the top of some natural hierarchy we imagine ourselves occupying.

I think the greatest gift AI will give us once we appreciate all the myriad ways in which it can be used, especially in creative endeavours, will be to show us the boundaries of the ways in which we are accustomed to thinking. In science for instance, science is often like a game of Jeopardy!, wherein, we have some observation or ‘answer’, and we are trying to formulate the right question to which that observation is the answer. And in devising experiments we are in effect building and testing differing models against one another to see which most accurately captures our observations.

In many ways then science teaches us about our own models and the constraints and boundaries they contain, rather than about the messy chaotic natural world with its panoply of causal arrows which lie beyond our comprehension. AI has made and aided in making tremendous scientific discoveries via pattern recognition in massive quantities of data that our brains simply have not evolved to recognize.

This kind of symbiotic relationship is really the hallmark of what AI can be, and in the field of art the only limit will be one’s imagination and the physical laws of the universe. I think the fear of job replacement - even in the arts is perhaps valid to some degree depending on the specific field, but I think the space of possibilities is so vast that any fears, at least for me, are eclipsed by the excitement of this burgeoning relationship with our silicon creations.

For interested readers, what are books, websites, articles or other sources of information you recommend for them to educate themselves on the topic?

These are just a few that I have really enjoyed, all non-fiction - some requiring a bit more persistence to grasp and finish - but all completely worth the time.

To Be a Machine by Marc O’Connell - easy to read and great little primer on much of what has been discussed in this interview.

Being You by Anil Seth - Though not overtly about AI per se, there is so much rich knowledge here about what constitutes a mind - probably my second favorite on this list after From Bacteria to Bach.

Ways of Being by James Bridle - One of the most fascinating, alarming, and also inspiring books about AI in various fields, including art. Also, really does a wonderful job of highlighting our myopic view of intelligence.

From Bacteria to Bach by Daniel Dennett - This is one of the real heavy hitters on the list, from one of my favourite philosophers - there’s so much fascinating multi-disciplinary material in here that the reader is simply left in awe of existence itself.

Possible Minds: 25 Ways of Looking at AI - 25 short chapters each written by a different author that reads more like the best sci-fi and philosophy you’ve ever read, rather than different non-fiction lenses through which to view our silicon creations.  

Life 3.0 by Max Tegmark - MIT professor’s phenomenal deep dive into the possible futures our invention of AI could hold in store for us. A must-read for those really interested in the topic.

The Meme Machine by Susan Blackmore - How do ideas propagate in a manner similar to evolution by natural selection and how does such cultural evolution lead to further feedback loops? This book will change the way you look at the world of ideas around you.

I am a Strange Loop by Douglas Hofstadter - For those who really want a mental workout and aren’t afraid to delve into the concept that what we call the self is simply an illusion - and how we discount our silicon inventions as mindless at our own potential peril.

I’ll add one final suggestion, the only fictional work on the list - a collection of short stories by the one and only Ted Chiang, probably best known for writing the short story that the film Arrival was based on. His collection is called Exhalation, and I’d be doing an injustice by trying to praise it with my own words here - needless to say, he, as a human being, is a gift to our collective imagination.


Previous page:
Part 1  
2 / 2
previous