I often think about art, and people who make art. And about democracy. But never have these two interacted more than when we talk about OpenAI. AI has slowly become an undeniable part of life; even Google searches are not searches anymore– they are AI-generated, guided responses that make your life easier, and set fire to a few trees. But that is not the ‘harmful’ AI, the one that simply answers your questions, the one that has truly captivated people, is generative AI, the one that ‘creates’. It is creative, and dynamic– writing personalized papers, literature reviews, articles, headlines, and copyediting ads. Now, it can even create artwork. All at one command. One sentence, one input picture, and voila you’re a creator.
Of course, as soon as generative art hit the market, people have been debating its use case. People have been divided. To the untrained eye, that art is just that, art. They couldn’t tell the difference between a painting made by Picasso and one generated by an AI. To be fair, who cares right? If it appeals to the eye it must be art, wasn’t that what the contemporary art movement was all about? Others have mulled over the loss of a whole labor market– artists and graphic designers losing work because of how easy it is to produce mechanized versions of artwork. Some hail this as democratization, why should boys(artists) have all the fun, am I right? I should be able to create quality art in the blink of an eye. Everyone should have access to the creation of art. And then there are those, that have declared AI as slop– lazy work, by a lazy LLM that does nothing but standardize art to the point of it being boring.
Before delving into any side of the debate, I want to take a step back. I have a friend. He might not consider himself a conventional artist, but he is a photographer. He takes pictures as a hobby. They are structurally beautiful and pleasing to the eye; the sunlight softly hits the flowers. But his lens is what is most unique about him. He captures movement in his pictures– slightly out-of-focus birds, kids running across a park chasing ducks, a car zipping past on a dimly lit street. The pictures have energy, and since I know him, I know that he does too. The pictures might not necessarily feature him as a subject, but they feature him. In the background of every frame, he is there deciding to capture the world the way he sees it, and it is this creation of the world, that makes his photographs exciting.
I have another friend. She is a writer. A caustic reviewer, a hard critic, her salty humor runs through everything she writes. Her writing runs through you, making you jump from one thought to another, following her as she takes you through a book, an idea, or a story. Her writing elicits a response; good or bad that is up to you depending on whether you agree with her. But her gamut of experiences and her own way of reading the world shows up.
During my undergrad, I did several translation workshops. Language I learnt, can convey meaning across cultures, but translation is not a strictly mechanical exercise. Many things are understandable but not easily translatable even if they have direct translations available. Sometimes the world of language that is inhabited by the text is not the same as that of the translator and might even be further away from the world of the reader. What do you do in that case? Translators make executive decisions according to the world that they understand, and the language they inhabit. I experienced language as a dynamic, cultural process where I wasn’t necessarily just writing the text in another language, but I was reinterpreting and analyzing a text, willing it to mean something. This is why a singular text can have multiple translations even in the same language. Each author writes into the text a world of their own.
These are all, creations, generated by human intervention. What all my friends will tell you is that, it took them several iterations, several hits and misses to reach the point they did. They will probably tell you that their process was not so straightforward, that they learned how to be a good photographer or writer not through just strict technical training but through learning how to articulate that form in ways that appeal to their sensibilities. Sometimes they might not even have a process, and it is only retroactively that they map out how they reached their final output. The most lively thing about their art, is it is full of mistakes. They have run-on sentences, shaky photographs, dull colors, and weird placement of commas. Their work has a rhythm that is undeniably human; a thing that AI standardizes and kills.
One of the bigger assertions that people make is that AI has made art accessible and democratic. Anyone(with an internet connection and a phone) can do it. But is AI democratic? Is there a certain privileged logic to creation? Should everyone be able to create? The answer is undeniably, yes. But what cannot be mediated or decided, is what should be the quality of this creation, or how much effort goes into this creation. Strictly on the level of assumption, where people with access to AI are also simultaneously people who know how to use AI– therefore– making them able to access ways of knowledge of such creation of art, real barriers to accessibility don’t exist. What that means is, that there are always communities that are often not afforded the time, space, or privilege to create, but people with access to AI do not fall under this. It isn’t that art isn’t accessible or that it is less democratic, it is that it is hard to create and requires effort. The relationship between democratic notions of accessibility and the creation of art might need a more serious philosophical inquiry.
But what I am more interested in is, is it really a creation if you are not working on it. If you aren’t going through the motions of writing every line, is it created? No, even generative AI, generates. Generate means to cause, to arise, to come about. The use case of generate is often for, energy– something uncountable, vague, and immeasurable– empty of intent. Create means to bring something into existence. You create art, God created this world– there was something to be contributed to, something tangible. If an LLM is generating ‘art’, is it making it really making it democratic, or just unimaginative, sterile, and uninspired?
A related question and something tech people often ask is, does art matter? I am biased in this question. I have lived a life inspired by evocative art– by the books I read, the movies I watch, and the music I listen to. I also aspire to live a life that is surrounded by people who do the same. I don’t know if art gives meaning to my life, but what I do know is that it makes it colorful, sharp, and fun. Whether or not I can make a living writing articles on my scroll stack is a question that is tangential to whether art matters. I have the privilege of believing that despite not having strict monetary value, art creates a world of its own and that is something that provides value in itself.
Is AI producing slop? AI writing is boring. It is technical in a way a textbook teaches you English– technically sound, but not fun. Poetry is fun, writing bad sentences is fun. AI doesn’t write up every word with intent care and interest– instead what it cares about is efficiency, precision, and an economy of words. Only useful when you need to write your 300th cover letter of the month. My friend spent several months working with me on my PhD applications, going over each line hundreds of times– decoding every word for what it meant. That is precision of thought, a technique AI cannot fully capture because it is only through human interaction, and editing that writing remains stylistically exciting and does not become like every other written piece on this planet.
AI writing makes us bad thinkers. Writing drafts over and over again, perhaps even bad outputs is what creates writers. Putting in the effort and time to read the whole text, instead of AI-generated summaries(however accurate they may be) creates memory. Writing is not just structured words on a paper instead it is a processual event. You think through the process of writing, the structure of your ideas, the things that go into the material the questions that come up. People will tell you academic writing is formulaic. It is structured, so maybe AI can do that better. Academic writing is formulaic in certain ways, but in other ways, it is unique. When you think about a literature review, the kind of literature you want to review, the gaps your research fills, hell even the way you structure that particular part of your writing– all become integral to your larger argument. These minute, but creative decisions are what makes us better thinkers– more critical and careful. It just makes us slow and less efficient.
The question we should all really be asking is, is art supposed to be efficient?
Write a comment ...