str(AI)nge love

Or how I learned to love the recent advances in AI-generated video in the art world.

“What made you choose this profession?” Whenever I’m asked this, I really have to fight my first impulse to wax poetic and regale any who would listen to the story of how I actually signed up for the wrong course in college. I didn’t have the heart to tell the lecturers I had made a mistake – they were too overjoyed to have more students than they had initially imagined. But if I am being honest, the course my professional career took can all be traced back to my experiences during primary school.

Every break time, a group of us kids would partake in a fun ‘round table’ of imaginary games. We’d play pretend and imagine ourselves as Power Rangers forced to battle each other in an oral recreation of an action-packed battle. However, throughout all this gameplay I had a unique problem, one that vexed me to no end. In each session, I would struggle to describe my intentions to my schoolmates. I remember the sheer frustration and it kept building with each iteration of our play fights. Until one day I just sat down right after our break, pulled out a pen and paper, and began to illustrate the visuals that were struggling to find purchase in my mind. I had no idea at that point, what I would choose as a career path, but I still remember the joy of breathing life into what became a tangible idea. One that no longer had to exist only in my imagination. It has been many years since those nostalgic days and that joy is being steadily overshadowed by the slow and inevitable creep of recent technology.

It has been eight years since I started in the content creation industry. At the time of writing this, I am currently working at Fuelcontent – a branded content creation company that specializes in live-action and animation. The job? Try and take an agency’s concept and turn it into a visual story(1). This, at times, means most of the effort is expended in the craft by taking the idea from being just a short sequence of text and refining it into a visual metaphor. As much as it can be fun, it can also be a time-consuming process. Despite clever solutions to ever-shrinking deadlines, the time requirements and technical aspects involved in its creation can be quite the behemoth. So, when the promise of new technology comes with the promise of simplifying and streamlining things, I always tend to give it the dirty side-eye. I am not gonna lie, I am a bit of an old curmudgeon when it comes to changing with the times.

When it comes to the recent technological advances within the field of Artificial Intelligence, I feel like this meme encapsulates the entirety of my belief..

Alex Blechman captures the spirit of modern tech companies (2)

Jokes aside, I am terrified by the steady encroaching pace of ‘AI-generated art’ and with headlines such as “AI generated art wins Colorado State Fair’s annual art competition” (3), “Horny AI Artists are trying to create the perfect woman” (4) and “AI generated Art Exhibition hits Amsterdam, fueling controversy” (5), can you blame me? All these and more articles are circulating each month. So, like it or not, this is something we need to talk about. But, before we go down this rabbit hole, allow me to set the stage and break down the definitions that’ll allow us to navigate the stormy seas of this article.

It was only a month ago that I was extolling the negative virtues of AI-generated art to my father, that my old man asked an integral question, “what is AI? Is it a robot?” I had to admit I had no idea at the time what AI was or why it was referred to by the term. Generative AI or AI, in short, is a term that refers to computer algorithms that try to understand a dataset of structures from existing content, such as text, audio, video, image files, and even code with the goal of generating similar or new content. In simpler words, it is an algorithm that tries to generate customized content according to a pre-existing data set input by a user. The AI must undergo a process supervised by a human that informs the algorithm through trial and error if its various attempts are the ‘correct’ result. The AI is limited in its ability to create content as it can only create from the data it has been provided (6). I am paraphrasing here but as I assured the old man, there is not much of a chance that the robot revolution will be brought upon by a squad of disenfranchised revolutionary robot artists.

9gag, says it best(7)

We should ask another question: what is art? Art, as defined by the Oxford Dictionary, is “the expression or application of human creative skill and imagination, typically human in forms such as painting or sculpture, producing works to be appreciated primarily for their beauty or emotional power.” (8) it is within this definition that the seeds of controversy lie. How does one define art created by a machine whose sole goal is to imitate without a fundamental comprehension of the emotive context behind the style or concept of the original piece? Also, if the art is being supervised by a human, is it truly AI art? For now, I am just glad we got the big scary terms out of the way.

It wasn’t too long ago that I was joyfully exploring the potential use of Dalle-mini. It was a hilarious experience of inputting commands such as “trail-cam footage of an American President” and getting a blob-like shape sans any defining features.

“Trail-cam footage of an American president”, downloaded from Dalle-mini.

And, one of my favourites, “Bearded man Slaps Uranus.”

“Bearded Man slaps Uranus” results from Dalle - 2

But as much as it was joyful, the experience can also be frustrating and painful. Seems like our robot overlords have not been programmed with a sense of humour or, at the very least, a fertile imagination. The AI would come close to the results I wanted, but perfection was still so far away (ah, perfectionism – the true nemesis of art). But plunge on I did as I figured now would be as good a time as ever to explore this new form of generated art. Yet, the more I did, the more I felt a need to understand why I was having such a visceral, fearful reaction to this new technology. It left me filled with utter trepidation.

?But before we get there, there’s a final hurdle of definitions I need us to deal with: Discord. What is a Discord? Discord is a social messaging platform that allows users the ability to communicate with one another much akin to Teams, Skype, and Zoom. Within Discord there are applications called ‘bots’ you can install to help the user enhance their own personal user experience(9). For instance, if I wanted to broadcast the 80s and 90s classics for my friends and me to listen to, as an administrator on Discord, I could download a bot and input specific commands as well as a URL link into the chat and boom – you have the sound of music playing for all those who have been invited to said chat room.

Within Discord, there’s an image generator called MidJourney that exists as a ‘bot’. Using a simple command prompt, with. /imagine a user inputs to the AI.

/Imagine a ‘Wes Anderson style interior of a bookstore’ and the results speak for themselves. Each low-resolution iteration took at most a minute to generate.

“Chibre Van Gogh style cinematic lighting” created by MidJourney

Personally, I have only just begun to dip my toes in this new medium. But trust me when I say there is a steep learning curve that forms the basis of understanding the text. I spent two hours waddling in the ‘kiddie pool’ as I flung text into the chat that would garner no results. I would have been lost and I would have probably struggled to progress any further if it wasn’t for the fact that MidJourney is a chat room. This allows its users to watch others navigate the AI. Watching others successfully and unsuccessfully send out prompts helped me grasp an understanding of how to phrase my requests in a way that generated appealing results. I would try to mimic their text prompts and then alter them slightly to see what would happen. It was during this process of watching and learning that I realised – I had spent hours creating set pieces. It was a singular rabbit hole of my own creation where I would experiment with ways to improve lighting and staging as illustrated by the images above.

It finally dawned on me; just how versatile AI-generated art could be. How concept work, which could take days of research, from concept to execution, could be accomplished in less than a minute. I was sold.

It wasn’t just me who was convinced by the creative possibilities that AI could bring. During my investigation, I brought up the possibilities of AI-generated art with a few friends and colleagues that work in the art industry. According to a close colleague and former lecturer of mine from my college years, MidJourney is already a part of the coursework that he teaches his students at Uppsala University in Sweden. His course covers how to use MidJourney as a tool for concept generation for young soon-to-be artists in the industry. This approach seems to be the general feedback I have received after talking to various professionals in the art world. The few who hold negative reactions to the tech are those who have yet to interface with the technology and have no intention to explore what is possible with it.

There is a pervading fear that AI will commandeer the arts and bankrupt the last few surviving artists looms over the horizon and articles that predict this outcome are only getting direr. Judging by the headlines, the slow encroachment of AI-generated art will not stop with generated images. Steam is slowly seeing a trickle of AI-generated games(10) and, as of last month, Meta and Google revealed their new text to video AI generator(11)(12). As more and more technology storms into the creative content creation space, it seems unlikely that it will stop. But is this the end of the arts? I honestly do not believe so. In my very quick perusal of this technology, I found it had the potential to be an incredible tool that could be invaluable to content creation. I can see it being used to create mood boards, generating concept art that could help directors with staging, lighting, and composing scenes that can be used at pitches, and aid directors when creating stages that bring their creative vision to life.

I won’t say what I have written is the gospel truth going forward. As a person who was initially excited about the creation of NFTs before they became a late-stage capitalist dystopia, I can honestly say, I have been wrong before. But I believe in the initial promise that AI-generated art, videos, and games can bring something of value to the table. It is not the art killer as sensationalised news would have you believe, for it still falls short of understanding the intention of the user and requires very specific wording to generate appealing art (and even that is a hit or a miss.)

Professionally, as a member of a content company in an industry where the demands are only growing, and the budgets and amount of time you have shrinking, I see the potential in AI. It can be a convenient asset. A tool that could cut down the time spent conceptualising, perhaps even affording us more time to refine our ideas and truly make them shine.