The first introduction I had to the world of art was an old surrealist book my mother kept from her art course before she had me. The book was challenging for a young kid to wrap his head around. Up until then I had been drawing goofy characters from Scooby-Doo that would end up on the fridge for a week or two. To be exposed to Salvador Dali and Juan Miro made me question at a young age why we make art. Far from the comforting arcs of Saturday morning cartoons, these surreal dreamscapes would frighten the viewer with a sense of bleakness and meaninglessness. This was to my memory the first significant challenge to my views on art.
Ever the empath, my mother told me that they were expressing how they feel inside. That sometimes if you feel like things don’t make sense, then we should share that feeling with someone else. Maybe that way we can come to an understanding on how to make it better. Later in life I was in my final year of college. I was studying fine art. During contextual studies we were posed with the following question
“What is Art?”
Fairly simple. Art is art obviously?
The point in this question of course, was that art is subjective. Everybody's answer is as wrong as it is right. It’s up to the individual to choose how they enjoy art, whether it be purely aesthetic, or deep analysis. Art can even be read in ways the artist never intended, and this would still be a fair interpretation of its meaning.
My answer was somewhere along the lines of “Art is expression of thoughts or feelings in whichever way the artist sees fit”
It was a good definition for me, as it didn’t exclude anyone. From Botticelli to my scooby doo paintings, to Jackson pollock. Every artist I could think of was expressing something. Be it a sunset, a moment in religious canon or a consumer product; every subject, presented in every medium was an act of expression of one thing or another and was thus art.
Notice that my own definition had the word "artist" in it.
The reason I started this with an anecdote is to illustrate my motivations towards this subject matter. That my world view is constantly challenged by how much art has changed since I gave that answer. With the advent of machine learning and "AI", images dreamed up entirely from advanced neural networks are making their way into the art space. If a machine can make art, is it an artist?
This idea first presented itself to me as short videos on my Facebook feed showing a mechanical arm holding a graphite pencil, which would “draw” various portraits and scenes.
The videos often came accompanied with some sort of clickbait title along the lines of “IS THIS IT FOR ARTISTS?!?!?!” "IS THE HUMAN RACE DOOMED?!?!?"
to both of which the answer was a resounding "no".
Of course, this “mechanical artist” was nothing more than a fancy printer, creating whatever the human artist told it to. It wasn’t until I was exposed to tools such as Adobe's "content aware" tools and “AI” image interpretation tools such as http://nvidia-research-mingyuliu.com/gaugan/ that the idea of a machine being able to crate imagery through recognising patterns became more prevalent.
The following twitter page shows an artist who uses machine learning to generate images from scratch https://twitter.com/images_ai?s=21
These loose, abstract, yet familiar scenes can often feel full of human emotion. Some evoke fear, others, a sense of nostalgia. These end products are the result of huge datasets, chosen and curated by an artist or programmer, which are then broken down into a set of sliders that denote different aesthetic decisions (often in a quite nonsensical manner). The process works in generations, meaning each step is closer to the final piece. I recommend trying https://www.artbreeder.com/ for an idea of how the technology processes image making. It’s a kind of alchemical mixing process.
With these sorts of network, a symbiotic relationship between the artists and the networks arises. The artist acts as disc jockey, remixing the collective artistic decisions of those whose work is in the dataset with the power of a neural network forming what it considers to be a sum of all parts. While the dichotomy between artist and machine is present, another party is involved. Each resulting piece of art is seemingly 25% artist, 25% machine and 50% the art of those in the dataset. This “standing on the shoulders of giants” element of the process is also true in traditional image making too – everything a person makes is a combination of things they have seen previously.
The problem is, a human being will replicate what they have seen imperfectly. Each artist works through their own ego, creating variety in works between individuals. A machine has near perfect recollection, and no sense of self. Two machines given the same data and set on the same task would invariably create the same image. Do we have a right to use the collective aesthetic decisions of artists in this manner?
If we look to the future, it is interesting to wonder if the collective artwork of human history would be enough basis to express a self-aware machine’s innermost thoughts.
These were among the questions I want to ask with my work, but mainly my focus was "where does technological advancement leave artists?" Or more "where does it take them?"
As Henry Ford demonstrated, seemingly sophisticated, expensive, outlandish technology can become an everyday piece of consumer product if it makes life easier. And it can completely change the world as we understand it in the process.
The following article talks about a piece of ‘AI’ art that sold for $432,500 https://www.christies.com/features/A-collaboration-between-two-artists-one-human-one-a-machine-9332-1.aspx?sc_lang=en
I want to draw attention to this extract from Dr Ahmed Elgammal
“There is a human in the loop, asking questions, and the machine is giving answers. That whole thing is the art, not just the picture that comes out at the end. You could say that at this point it is a collaboration between two artists — one human, one a machine. And that leads me to think about the future in which AI will become a new medium for art.”
What Dr Ahmed Elgammal illustrates is that a machine can use input information to create images, but a machine doesn’t have a “self” to express. The empathic responses of images created with this process are from the human being. Much like a director and his crew, the human being leads the network towards a satisfying final product.
In the extract He states, “The whole thing is the art”. I find this to be a good insight, as it highlights that we as an audience seek not just a final product, but some sort of narrative involved with the process.
With this subject in mind, I wanted to present “Immersive Arts” as art and the artist immersed in digital collaboration.
While true artificial intelligence is still in its primordial soup, and beyond my understanding, I sook to create work that ask these sorts of questions through imagery. Interpretation and observation are some of the most foundational forms of artistic expression. Artists such as van Gogh and Picasso, have particularly unique or nonconventional methods of representing the world around them and are often seen as visionary. My interest is in the ways in which a machine’s nonconventional expressions of reality can give the machine the appearance of a true artistic collaborator. The project changed from the initial proposal, but what I kept was the idea of “Rhopogrammetry”, A portmantua of photogrammetry, the 3D reconstruction of scenes through photography, and Rhopography, the study of everyday clutter.
the reason I chose everyday things as the subject is that it comes from a long tradition of artists. From Cézanne’s still life to Marcel Duchamp's ready-mades. The world around you tells a story. Ideas of form and function are deeply connected to our understanding of every day objects. This is a comfortable ground to experiment from, as it is quite low profile as “artistic voice” goes. I wanted the machine to talk louder than I did so to speak.
As for the machine, photogrammetry using 3DF Zephyr was an important starting point. If you read my previous entries, you will know that photogrammetry is incredibly difficult to get right with a basic setup. I'm hoping that I can find some sort of intrigue from those errors. the mistakes and imperfections of digital reconstruction show the machine’s unique world view
My focus then grew on a series of experiments that aimed to blend the machine’s imperfect view of reality with my own.
My first port of call was collecting various photo scans of things around my apartment. the first was a carefully assembled still life. the resulting photo scan - admittedly - was a mess, however it wasn't my place to tell the machine how to "express itself".
ok I messed up the scan
Its roughness and bumpiness would have generally been avoided by a painter or sculptor, but I would argue that my human error pushes a more distorted, digital result. The interesting thing with a machine recreating an image as apposed to a human being is that a machine doesn't get caught up in symbology. where a human being might assemble this image by clearly distinguishing the "skull" along with all its meaning and subconscious connotations, the machine simply arranges polygons in a mathematical manner and projects photos onto it. the result can be a sludgy, rocky, surreal equivalent. The
while working on this particular experiment, I found a software by the name of Ebsynth. it seems to marry artist and machine perfectly, using video codec logic to warp custom frames to the motion of videos. I created a turntable animation of the underlying 3D and created the following frames to overlay
these frames provided a painterly style, as well as the more human notions of meaning, separate objects and symbology. when fed through the software, the median frames were generated creating the following effect
decisions I never would have made can be attributed to none other than the computer and its imperfect approximations. Of course, these systems were created by humans - so ultimately it's human error, however, this was never anybody's intention. If this project is a metaphorical, mythological journey into the artist inside the machine, then I can think of no better expression of this.
While I liked this experiment, It was carefully arranged, and later painted by me. The top layer is human art, i wanted the machine to talk first.
I often hear photogrammetry referred to as "painterly", so the next step for me was to allow the machine's "style" come through in some way, so I set out to create a piece involving my desk. With these subsequent experiments i tried not to interfere with the space in the way it was presented. The setup here was how i had it set up for working. I had a feeling that if i start to interfere with the arrangement in order to create something artistic, it draws the spotlight on me not the machine.
One thing to note is that the computer that constructed this model in on the desk. Could this be considered a self portrait?
For good measure I also scanned some parts of me and my partner's messy bedroom as an homage to Tracey Emin.
What stands out to me the most is what the software decides to do when the information isn't so clear. where it doesn't have an angle to project information so it just makes something up or cuts it out entirely. i also like viewing these models from inside, every extrusion becomes inverted, convex become concave and vice versa. the results often challenge our perception of form and function.
When i took this into ebsynth, rather than painting over the machines interpretation, I used the machine's interpretation as the top layer and added some edited footage as the bottom layer. the effect crated is almost psychedelic, the whole world shifts around trying to catch up with itself, things phase in and out of existence. its interesting to see how the layers of the computer's mistakes stack up the feeling of irrationality. For good measure I also slowed the footage and allowed after effects to digitally interpolate between frames.
I call this piece "using the computer while under the influence of narcotic substances"
my place in this piece is still obvious. the recursive loop is intended to appeal to a human viewer and create meaning and rhythm. What I refer to as the computer's mistakes may still be my mistakes. 2 machines with this task would almost certainly would have come to the same conclusion, however another human would have produced different errors.
during my Rhopogrammetric (trademark pending) experimentation, I did wonder what it would be like to put another ghost in the shell, so to speak. Working on human subjects for handheld photogrammetry has been a mixed experience to say the least. in my previous entries I mention how still things are much easier to work with. Slight differences in eye position, or overall stance between shots are natural in people. Again i am going to use this as the machine's voice.
the first human subject was my partner who has also featured in my previous photogrammetry work.
This has its errors, but this is the happiest I have been with a human scan so far. I'm shocked this isn't a more popular thing to do. I could imagine someone having a 3D scan portrait of their grandma in her 20s lying around on the family supercomputer. Or how about 3D scans of important historical figures to refer back to. A portrait gains dimension when processed this way.
The errors in this image also demonstrate the computer's inability to symbolise with this process. the hair meshes with the cheek and neck in a disturbing way. Of course - to the machine "disturbing" was never the intent, this is no different to the strange artefacts in the still lives. I find it interesting that when it's on a human, tech distortion is much more uncomfortable to us.
speaking of which...
this was incredibly difficult as it was entirely made out of "selfies", meaning my body had to move in order to get the different angles, but the resulting digital confusion present is certainly what I was looking for.
In the spirit of the previous ebsynth experiment, I relit it to highlight the imperfections in the geometry as an overlay and used footage of myself as a basis. the resulting video is below has a painterly feel that is completely generated by a machine. there is not a single brushstroke in sight, to say it is my style would be a mistake. so the question i want to ask is this
Is it a self portrait, or a portrait?
I'd like to say that it is a portrait of me by a machine, however any attempt on my part to do so would be metaphorical. To imply any true authorship onto the machine in a this early stage is in the realm of science fiction. Every key decision made by me was in order to create some empathic response in the viewer. That is my human ego at play. even the errors I highlighted as works of the machine were left in, drawn attention to, and assigned meaning by me.
Yet in a similar way to the neural network image generators, every human artist has a visual library that they borrow from. thought and behaviour are reactions to information taken in by our senses. We are machines of our environment, creating art to appeal to each others unique coding.
As we look to the future it's hard not to wonder what the world of image making will become, especially with the economic boom of NFT art giving AI developers a financial incentive for rapid image production.
I hope this has been insightful into the area of work that has held my attention for these past months. Art for me is about empathy, while machines are far from a sense of self comparable to ours, it still makes me wonder...
…what will the first truly artificially intelligent artist have to say about us?