What Is Art?

We all know what art is, right? Or maybe not.

Art used to be fairly easy to define. It’s kind of like the accepted definition for pornography. You know it when you see it. But as technologies change, it becomes harder and harder to come up with a universally accepted definition of what is art and what isn’t. And, to take that a step further, who is an artist? And who isn’t?

You don’t have to be a particular level of genius to figure out that this thought is bouncing around my mind at the moment largely as a result of the ever encroaching force of artificial intelligence. This essay won’t be specifically about A.I. I’ve written at length about it. The executive summary being that it is a massive force that will bring a seismic shift to our industry, a currently underestimated threat to intellectual property rights, and a tool that will put a large percentage of artists out of work. If that last sentence seems particularly dramatic, I assure you, after taking the last two months to dive deeper and try to understand the technology myself, I feel as though that level of worry may still be sugar coating things a bit. But, regardless of any reservation I may have, A.I. is an undeniable coming force. It’s a genie with no intention of going back in the bottle. Along with the clear negatives I pointed to above, there are also market drivers which will ensure its rapid acceptance. We can choose how or if we are going to engage with it. But it will, without question, engage with us, our business, and our art form whether we want it to or not.

Just consider the number of photo competitions this year being compelled to address whether or not they will accept A.I.-generated submissions. I have a clear thought on this. To me, a photo is something taken with a camera by a photographer. An A.I. generated image, no matter how photorealistic, is not a photograph. And the person who created it, is not a photographer (at least not with regards to that particular A.I. generated image). What I’m less clear on, is whether or not this is ultimately going to matter to future generations. And if we accept the fact that A.I. “art” will exist, how will it co-exist with existing art forms like photography and filmmaking?

If you’ve read any of my previous articles, you will have no doubt assumed me to be a grumpy old man averse to change. You would be right. Things like social media, selfies, and whatever you call those darn scooter things young people ride down the sidewalk when I’m just trying to have a pleasant afternoon walk, all get my grumpy old man sensors triggered. But the other thing that comes with age is perspective. So I know that, while those things may drive me up a wall, I’m well aware that many of the things that I did when I was younger drove my parents up a wall. And many of the things they did drove my grandparents up a wall. And so on. Times change. Norms change. And life goes on.

As does art. As someone who takes their photography very seriously, and as a student of history, I can never fail to remember that there was a time in the world where people didn’t consider photography to be a serious art form. Painting was the real art form. Photography, created partially with the aid of a magic mechanical box, wasn’t considered art in the same way. The machine itself was seen as an unfair advantage. As time went on and photography established itself as an art form, it started to replace the use of painting and drawing in certain situations. It became its own thing. And, its practitioners found ways to separate themselves out as artists regardless of everyone having access to the same tool. Painters and sketch artists didn’t cease to exist. But the market shifted and some tasks that previously would have required a drawing, for example, could now be replaced by a photo shoot.

One of the biggest legitimate complaints about A.I.-generated imagery right now is that the machine is doing all the work. You add the prompt. But you don’t actually have to know anything about the exposure triangle, lighting, or lens choice to create a perfectly acceptable photographic (feeling) image.Having now been futzing with A.I. for a couple of months, I can say that understanding lighting and knowing how to describe what you want does make a difference. So, actual photographers using A.I. systems do have a built-in advantage. But, it is also completely possible that a total novice might still generate a perfectly lit image by accident simply because the A.I. made the decision for them. Since a large part of being a professional photographer is pairing your artistic voice with in-depth technical knowledge only gained through years of hard work and experience, it seems unfair that someone without any level of technical expertise can have their work considered in the same breath just by typing in a set of keywords and hoping for the best. But that is part of the new reality.

Worse yet, while you and I might have the integrity to create keywords wholly from our own imagination, let’s not fool ourselves into thinking that the vast majority of people who ultimately use these platforms won’t stop simply at “give me an image of this in the style of (name your favorite artist).” This goes beyond using an artist for inspiration and, in my opinion, is literally theft. The other day, as purely an experiment, I wanted to see just how easy it would be to ape another photographer’s style. I typed in a series of prompts for different characters all to be shot in the style of Annie Leibovitz. And, in a matter of seconds, one after another, the computer created spot-on images of characters lit, styled, and posed almost exactly as you’d imagine Annie would light them for a spread in Vanity Fair. If you didn’t look too closely, you’d be excused for thinking these images actually were shot by Annie Leibovitz.

I’m not going to show you these images because, to me, this is not art. This is mathematical imitation. Nothing about the images or my contribution to them makes them feel like they are mine. I did this exercise purely as a private experiment to see how easy it could be done, going about it in a way that I suspect much of the public will use the technology. And sadly, it was very easy. Yet still, the images, while accurate, feel as though they belong to Annie and all her hard work of creating a specific style. So, no matter what I may or may not have typed into the computer, those images don’t feel as though they belong to me. But that’s me as an artist speaking. John or Jane Doe, who have no experience creating art or developing a unique artistic voice, could do the same thing and might very well feel ownership of the end result. They may even go so far as to consider themselves artists, even though the “art” they are creating is literally based on someone else’s intellectual property. It literally only exists because so much of Annie’s actual work has been used to train the machine, which, in turn, can regurgitate a fairly faithful facsimile of the real thing.

There’s also the question of aesthetics. In just the short time that I’ve been learning about A.I., the aesthetic capabilities of the systems have exploded.  What I can do now versus what I used to be able to do just two months ago is absolutely crazy. A common refrain is that A.I. generated images still don’t feel photographic enough. And, that may be the case. But I can’t help but to feel like that’s a problem with a solution in the very near future. As in a matter of months, not years.

But even that raises larger questions. When I began my career as a filmmaker and photographer, I was shooting on film. The transition for still photographers to digital was rapid. But the idea of shooting a major motion picture on digital video instead of film took some convincing. Early attempts delivered poor results. There just didn’t seem to be a way to capture the majesty of film production while shooting in digital formats. Of course, over time, technology developed and now film production is the one in the minority. More than that, audiences' acceptance of a more “digital look” versus a 35mm film look has changed. After years of digital becoming the standard, audiences now accept that a movie can be shot digitally and still be beautiful. You may still have the choice to make your movie on film, but it’s an aesthetic choice rather than a necessity.

I can’t help but to think about this in terms of A.I. The images produced by these systems do currently look different than what we’ve come to accept as photorealistic. However, as these systems become the norm and more and more young people grow up seeing A.I. generated faces everyday on social media, will we reach a point, like viewing movies shot on video, where the A.I. aesthetic just feels normal?

I think of superhero movies. I’m not a fan. Really, really not a fan. But they have overrun cineplexes for over two decades now. An entire generation has been raised with superhero movies being what they consider to be the theatrical experience. Everything else is just fodder for Netflix. So, while I personally might see a Marvel film as more of an amusement park ride than actual cinema, I have to accept that people younger than me will feel a completely different way. Not that either of us is wrong or right. But, it’s a matter of perspective. So, twenty years from now when we have a generation of young people who have grown up with A.I. as a reality driving the culture, might we reach a point where A.I. is the default instead of cutting-edge technology?

And what happens to artists in that world? How do we define an artist in a world where so much of the technical skill required to achieve a result is subtracted from the equation? If anybody can create something that could objectively be considered artistic just by punching in a couple keywords, then is the result still art? Or is it just data?

Now, I should say that one thing I have learned from experimenting with A.I. in the last couple of months is that it does matter what you put in as a prompt. There is a certain skill set required to create consistent results. In fact, one of the more interesting things I’ve noticed, as more and more of my fellow commercial photographers begin experimenting with A.I. image generation, is just how much one’s artistic voice still comes through. For instance, my photographic style is very cinematic and story based. This being a result of my background as a filmmaker. Predictably, the A.I. I have generated has all been very narrative in nature. But, on the other hand, I have a good friend who is a beauty and fashion photographer. And her experiments in A.I. have been very fashion oriented and look very similar to her particular brand of beauty photography. Another more conceptual photographer I know, creates very high concept A.I. work. Regardless of the tool being used, the artist’s voice seems to be shining through. Whether or not that matters in a world where someone without an artistic voice can simply steal yours by typing your name into their prompt is an entirely different story. But, this little survey of my fellow photographers’ work suggests that there is still a world where great artists (I’m talking about them, not me) can still set themselves apart through the unique way that they see the world.

The technical acumen that sets a professional artist apart from a hobbyist seems more in jeopardy. Being able to craft consistent prompts currently is one way to set yourself on a higher level. However, I suspect that we are not far off from 3rd party software (some already exists) where you don’t even need to learn how to write a prompt in the first place. Rather, you just click a series of multiple choice options and the software will automatically write prompts for you. So, differentiating oneself purely on prompt-writing may be more difficult than, for example, differentiating oneself based on decades of experience with lighting. Also, as a professional photographer, you’ve probably had the experience of having clients assume they can do your job simply because they have a fancy camera that can take pictures in Auto mode and they have convinced themselves that your job isn’t about skill, but just about having the best camera. Imagine how fun that discussion is going to be once clients feel like they are also perfectly capable of lying in bed typing random prompts into a computer and having it generate “art” that objectively hits the minimum threshold for professional quality without having to pay your fee. That’s a rant for another day.

But again, that all begs the question. What is art? And what is the role of the artist?

Is it merely a matter of technical execution? And, if so, what happens when the technology undergoes a sea change? If you transition from being a painter to a photographer, are you fundamentally a greater or lesser artist? If you transition from film photography to digital photography, thus gaining more access to digital tools, are you more or less of a photographer?

One of the things that always sets apart great art for me personally is the degree of difficulty. When I see an image and can’t for the life of me devise how someone was able to capture it, they get bonus points in my estimation. But, if A.I. can make some of that technical ability superfluous, then how do we accurately assess the artist's skill level? I would say the end result. But if that end result only exists because the machine has been trained on the hard work of another artist, how much credit does the second artist really deserve?

Or is it simply the idea that makes the artist? After all, all the technique in the world doesn’t amount to a hill of beans if it’s used to execute a subpar concept. But, then again, it can’t only be the idea. Take, for instance, a hypothetical example: a studio head gets an idea that he wants to make a love story set in Morocco. But that’s all he’s got. Then, Julius and Philip Epstein come in and craft a timeless screenplay that becomes Casablanca. Who’s the artist in that equation? The project wouldn’t have existed had Jack Warner not assigned it to the writers in the first place. But Warner didn’t know how to write screenplays, so there’s no way it would have been executed in that fashion without those specific writers. Who is the artist? And what of the original play on which the film is based? What percentage of the artistry belongs to source material?

You’ll notice that I’ve asked a lot of questions today rather than pretending to know all the answers. I’ve talked a lot about artificial intelligence because it is going to severely upend everything we know about art and the way we create it. But it doesn’t change the fundamental question. How do we define artistic ownership? How do we choose to value artistry when aesthetic considerations change wildly over time? Is our artistry bound by our specific artistic discipline? One thing that I’ve always found funny is how my screenwriting, cinematography, directing, and still photography (and now A.I.) all seem to feel like they come from the same place. The method of execution seems irrelevant. As an artist, I have things I want to say and the tools are merely conduits. But I do also highly value the amount of technical knowledge required to use each of those tools. And, if we live in a world where technical knowledge is no longer required, how does that enhance, or dilute, the art forms themselves?

Art, and the creation of it, is such a highly personal and subjective exercise that it can be difficult to accurately assess its value. Just the word “art” can take on a million different definitions depending on the person doing the describing. But one thing does seem to hold true from the beginning of time. Whether you are creating art with a quill, a paintbrush, a Nikon, or a computer, art is a constant tool for personal expression. And the more of you that you put into your work, the more special the end result can be.