The majority of posts on my social feeds over the past two weeks have been beautiful, fantastical portraits of my friends that were created by a combination of computer algorithms, after being fed a dozen source images of each friend.
These images are beautiful creations (with the occasional atrocity) executed in a variety of styles. That a computer created them programmatically is mind-blowing.
Magic in Three Dimensions
As a kid, the first computer I saw was monochrome green on black. In the span of 40 years, we’ve progressed to the near-divine: 3D gaming that increasingly approaches life-like, powerful computer vision applications in commerce and architecture, and uses programmatic means to interpret captured images.
We created the discipline of photogrammetry, using source images to create navigable 3D spaces. This was revolutionary a couple of years ago but this month photogrammetry is already in the rearview mirror and the new hotness is neural radiance fields. I live with one of the global leaders in that field and as a physics guy, watching and learning how they capture and recreate light fields to simulate reflections, and refractions is just next level. Once a scene is captured, a camera path can fly through the scene, creating an output video from a virtual camera. When shared online it can be hard to imagine that the video wasn’t actually captured in real life, but rather rendered.
This is the state of computer graphics and it’s astounding.
Technical, Sure… But Now It’s Making Art?
It’s one thing to analyze images, calculate angles, and create a technically-computed output world. Amazing as it is, it’s quite another to have algorithms that create what we consider to be art. Lines and angles, no problem. But now images are pretty, they have character. We’re witnessing an intelligent blend of existing styles and source material, with input provided via descriptive phrases or source images. A couple of such phrases might include “a cowboy wearing a tuxedo on the moon” or “horse galloping at sunset painting Andy Warhol style” (phrases lifted from an article on Medium.com)
I’m an art photographer with some design experience. I also make my living as a software engineer. It’s from this vantage point that I have to declare that these images are incredible and that the technology behind them is an amazing feat. When I see some of these images of my friends, I smile. These are not captures of reality but rather the gift of inserting someone into their fantasy dream.
The spirit of life is built on the power of story.
Me
Visually, I’ve never seen anything that can more easily cast a person into a story than these new AI tools. This is moving people in a powerful way and they don’t care as much about the process as the result. These images are fantasy, they are story, they are a dreamy escape.
The Real, The Magic, The In-Between
Editorial photographers capture the reality of a moment in front of them; for them to edit their work is considered unethical. Portrait photographers almost always retouch their work, removing blemishes, sometimes smoothing skin to the point of an image stepping into the fantasy realm. I’m a light painter, creating long exposures in the dark, and waving lights around to perform an image. We create something in between those bounds, and you can see my work at HackTheLight.com or on Instagram. Other than some color balance I choose to not edit my images, and yet the result is still something magical and unique; a portrait that is unlike the subject has ever seen of themselves.
I feel like I have definitely “arted” and earned my creation. Painters and illustrators feel the same, probably even more so.
A Novelty and Viral Moment
AI art is not going away; we have barely even started to see what is possible. This is both exciting and terrifying. Lensa AI Magic Avatar and Dall-E, both built on Stable Diffusion, are giving people something to get excited about, and they have even made the experience friendly for non-technical users. Opinions are myriad about the value of these images as “art” but even most of the detractors would have to admit the images are of high quality. One of the challenges is that Stable Diffusion dances through legal loopholes as it uses traditional work from traditional artists in order to even work, but without providing attribution or compensation to those same artists.
Suddenly everyone is a self-proclaimed artist, in that they performed an action and images were created. Is this what we have traditionally considered art? Nope, not even close. But tools evolve and the use of quality tools must be respected. Maybe a new phrase other than “artist” is in order. Saying a few words or uploading a source selfie is not the same as spending a month on a painting, or executing a challenging photo shoot. Still, images are being created and I love that this technology is instilling some version of the joy of creation into my friends, and the world. It’s something… but it’s something different.
Some of my artist friends are using generated AI images as a starting point and then using their personal training and skills to make them their own. I like that. Hate this or embrace it, it’s not going away. Seeing them adapt and try something new is encouraging. (Check out Mellow Monsters by Adriana Michima)
I tried out Lensa Magic Avatar in order to be able to speak to it with knowledge. The image accompanying this writing was from that sampling.
Feelings an Artist Might Be Left With Right Now
As for me, I’m tired. The over-saturation of AI images in media-space is getting to me.
The trigger for this article was in seeing the enthusiasm that one of my friends had for his outputs, and I don’t think that enthusiasm was misplaced. One of the images was perhaps the coolest artistic rendering of him I’ve seen. Yes, yes, I just called it art. My mind then flashed back to an image I worked hard to capture, navigating angles and lighting conditions, as I caught a powerful live photograph of him in the middle of a DJ set. At the time I would have considered that one of my favorite images I’d ever seen of him. Where AI art is now getting drools and spirit worship, that image got a single “like”. One like… nothing more. The enthusiasm for generative art has, by an order of magnitude or more, eclipsed the enthusiasm for more artist-produced work… at least this week. That’s hard on the spirit.
With all of the time, learning, practice, equipment, printing, framing, and more that goes into my work, seeing a few phrases with a few seconds of work create these new visuals, visuals that everyone has lost their mind about, it leaves me exhausted to think about creating anything more in the way I have. The idea of even doing another light painting photoshoot seems like a waste of time. Between my dogs, my dating and social life, and building my Multipass.com ticketing platform, I have to decide if I will continue to find it worth my time to invest in art as I have done. Right now I’m not sure I’m feeling it.
Maybe I’ll switch back to electronic music production, as I have been threatening to do now for 18 years, but music is probably next on the AI bandwagon. It’s probably already a lot farther progressed than I realize. I’ll be looking into this with fascination and I might report back on this.
Maybe I’ll make a nice mushroom tea, find my peace in the reality that truly is around me, and put thoughts to writing, sharing the unprogramming I’m achieving in my own life and wishing some of the same for those around me.
Maybe I’ll still make art. Perhaps I’ll pack up my gear and dust it off occasionally when I’m feeling nostalgic for the age of “art before robots”. Maybe it’s time for something else to move my spirit.