Generative AI and the art crafts
Hi there! The idea in this article is to give an opinion and open the debate about the impact and possible consequences of the growth that generative artificial intelligences had in the artistic trades in recent years, in the form of a disordered cascade of reflections. It comes out on March 1st and not on February 30th because February 30th doesn't exist and also because I took a vacation and couldn't make it in time. Sue me.
All the images in this article were generated with Midjourney using different prompts and with photos from my article "Travel light on gear" as input. I thought it would be fun to set the mood for the post 😊.
The already classic disclaimer, I am a photographer with a technical background from years of working in IT. I am not a specialist in any of the topics I'm addressing (I even think that sentence has a few too many words). All of this is opinion. Or a list of excuses written by a machine that inhabits my brain and pretends to be me, I don't know.
Now, after the playlist of complete albums from every fortnight and a full concert (the recording sessions of Snarky Puppy's Empire Central), let's get started.
Is generative AI a threat to the artist's craft?
First, note that I'm talking about "artistic crafts" and not "art" itself. Art, understood as a cultural activity driven by the desire or need to express or communicate a sensation, feeling, or vision of the world through various media, is completely safe. Humans will continue to express themselves artistically as we want and can, and consume expressions from other humans, as long as our species exists. We'll never escape from that and from stupidly killing each other.
Similarly, the emergence of new techniques and technologies doesn't kill off the old ones, although it may shift them out of the public spotlight. Books are still here (even if they're now ebooks), radios are still here (even if they're now podcasts), TV is still here (even if it's now livestreams). Painting didn't disappear when photography was invented... In fact, it was freed from the burden of trying to imitate reality perfectly and gained new possibilities as a result. We also didn't stop going to concerts or theater performances when a machine could reproduce them in our living rooms.
Focus is a bi-weekly newsletter written by Nacho Dramis. Subscribe to get it free in your email. If you enjoy the content and find it useful, you can make a financial contribution to support the project through Cafecito (Argentina) or PayPal (Worldwide). Making this content for free and ad-free takes a lot of time and effort!
Este newsletter también está disponible en español.
What is supposedly at risk and the reason for the generalized paranoia are "artistic crafts," people who offer their artistic skills and techniques as a service to others, such as writers, illustrators, cartoonists, architects, audio mixing and mastering engineers, some roles in photography and audiovisual production, and a couple of branches of design, among others. Over the past two or three years, there has been a flood of AI models, either freely accessible or very affordable, autonomous or incorporated into existing tools, that generate content of various types at astonishing speeds and with quality that continues to grow by leaps and bounds. In case anyone is unfamiliar with what I'm talking about, some of the better-known examples you can search for include text-generating AIs (ChatGPT, Bing/Copilot, Bard/Gemini, Claude), image-generating AIs (Stable Diffusion, Midjourney, DALL-E, Firefly), and even video-generating AIs (Kaiber, Runway).
Those who are undoubtedly losing the battle or have already lost it completely as I write this are those who produce "generic" or stock content, whether it's photography, audio, graphics, etc. Current generative AIs are perfect for that kind of thing (fueled by all the material generated by humans over the past few decades, a moral issue I'm not interested in debating here), and I can't think of why someone would buy a stock photo of, let's say, an apple on a table, or pay a photographer to take one for their exclusive use, instead of generating one with AI. The same goes for the hundreds of thousands of blogs with generic content designed solely to attract Google searches... So yes, 99% of the AI-generated content we find on the internet is inconsumable garbage created to try to make some advertising money from the distracted, but that was already happening before, and those producing it are the same people, they just have a more efficient method now... we're not going to blame machines for our bad practices.
Others who have been losing, against all odds, are the technology giants, drowning in their sea of bureaucracy and fears of traditionalist investors, being overtaken by startups with fluid capital. With the exception of Microsoft, which invested in OpenAI in time, beasts like Meta and Apple timidly stick their noses out without impressing anyone, and others like Adobe and Google leave much to be desired with Firefly and Bard/Gemini, respectively.
And that's talking about the present, whose distance to the distant future in these matters is only a couple of months away. We are very close to the technical feasibility of surreal and psychedelic situations such as weddings where there is no photographer, but the photos are not taken at the moment, they are generated later (or before, nobody cares) by an AI from existing photos of the couple and the guests. And here we find the first artistic trades that at least for now retain their value intact: Documentaries. The capture of something that happens, in the moment and place where it happens, with whatever means, remains for now irreplaceable.
Since I use the term "irreplaceable"... Amidst all the debate about the morality of asking an AI to "write this text in the style of Lovecraft" or "create a drawing like those of Akira Toriyama" and such specific prompts... What has always been truly irreplaceable about people in the world of art is their mind, not their hands. Toriyama (and all recognized artists from all fields and geographies) always had assistants for coloring, inking, etc., and the same applies to other mediums. Personally, what concerns me here is the factor of the "school" of the recognized artist with assistants, a practice that has existed at least since the Renaissance, and if these assistants are gradually being replaced by AI, then the vertical transmission from master to disciple of the craft begins to dissolve.
We reach the strongest point of this mishmash of words: Generative AI does not have judgment. Repeat after me: At least as of March 1, 2024, when I write these lines, generative AI does not have judgment. Generative AI still does not know how to design in the purest sense of the word; it only knows how to create beautiful things. It does not provide solutions to problems; it offers solutions in the air based on known patterns (and this is also a problem faced by many design professionals and artistic trades, but that is a topic for another article). The real magic lies in the idea behind it, and in what and how one asks the artificial intelligence, and how one continues to work with what it delivers. In the right hands, it can enhance a good idea, but it can never save a bad one.
It's a tool, and as such, those with more experience in the field will use it better, as it happened with the generation that moved from drawing on a desk sheet to using Photoshop, Illustrator, AutoCAD, etc. Like what happened to audio professionals who transitioned from magnetic tape to DAWs and plugins. Like what happened to photographers who moved from 35mm film to digital sensors, from the darkroom to the PC.
Even so, after the novelty boom passes, old techniques and technologies resurface and remain relevant, whether out of preference, nostalgia, or the superiority of the old format in some aspect over the new ones. In photography, analog photos are increasingly seen even in spaces where digital reigns comfortably, such as concerts or sports events. Transistorized amplifiers, digital pedalboards, and carbon fiber guitars have existed for decades, yet the vast majority of guitarists seek and go crazy for tubes and mahogany. Not to mention the revival of the warmth and imperfection of vinyl records over the cold perfection of digital formats when it comes to reproducing music.
We're on the brink of AI solving some issues perfectly. I believe it will be the imperfection and dedication of human beings, their ability to express what makes them unique, that will continue to make the difference... and, after the novelty fervor fades, be the aspiration to pursue. Like the incandescent heat of valves standing up to the zeros and ones.
Then what?
This newsletter is written in its Spanish version and then translated into English using AI. Do I have the knowledge of English to translate it myself? Yes, but it would take me ten times longer and the result would be inferior to if I dedicate myself to correcting details in what the AI provides. Am I depriving a translator of work? No, because if AI did not exist, I could not hire one either; I would do it as best I can. But well, topic for another debate, the non-artistic professions and trades that are losing ground with these technologies. Could an AI write an article similar to this one? Yes, of course, and it would probably be better written and all. But this is what I think, this is ME expressing myself, and you on the other side reading my worldview, as I said in the first paragraphs... I want to write it, and you want to read it, and no AI affects that.
AI can already make music too... Are you, as a musician, going to stop making it because of that? Are you going to stop listening to music made by humans? Me neither.
For now, I embrace it as what I believe it is for now: An incredible working tool that allows me to do more, better, faster, and with less effort, and a creative tool that opens up expression boundaries that were previously fenced behind technique.
Additional reading
Here are some articles and videos that I liked to delve deeper into the topic, some of which inspired mine.
Alva Majo, a Spanish indie game designer, on the ethical dilemma of using AI or hiring artists.
Some of what's coming up, OpenAI Sora, video generation from text, from the excellent Substack "The Algorithmic Bridge".
Another one from The Algorithmic Bridge, comparing the AI boom with the birth of the internet and thinking with monday’s newspaper in hand.
Anson Cheung and Héctor Rodríguez about AI and design.
Benj Edwards from ArsTechnica about accessibility in art with DALL-E 3.
https://arstechnica.com/information-technology/2023/11/from-toy-to-tool-dall-e-3-is-a-wake-up-call-for-visual-artists-and-the-rest-of-us/
An incredibly complex paper on the subject from late 2023. https://www.coatue.com/blog/perspective/ai-the-coming-revolution-2023
Hamish McKenzie about the value of human connection in the age of AI.
IA as a tool to empower knowledge areas you are already comfortable with, by Patrick Morgan.
AI and UX by Raoul Flaminzeanu.
https://uxdesign.cc/from-pencil-to-siri-the-evolving-function-of-ux-design-8ee23ef725ee
Nicholas Charriere’s 2024 AI predictions.
https://axflow.dev/blog/ai-predictions-2024
Focus is a bi-weekly newsletter written by Nacho Dramis. Subscribe to get it free in your email. If you enjoy the content and find it useful, you can make a financial contribution to support the project through Cafecito (Argentina) or PayPal (Worldwide). Making this content for free and ad-free takes a lot of time and effort!
Another way to support the project is by purchasing prints and various items with my photos in my international store on RedBubble. I don't have a store for Argentina yet, but I hope to solve that soon.
The links to Notion and Moment are affiliate links, which means that if you buy something from them, you pay the same and I earn a commission.
You can also follow me on Instagram, Behance, and Twitter. Sharing my articles helps a lot and is free :)
Este newsletter también está disponible en español.