Brady Walker: Can you tell me about your background in both technology and art?
Daniel Ambrosi: My first degree was in architecture and, while I was a sophomore at Cornell University in the spring of 1978, one of my professors gave a public lecture about what his lab was doing on the ground floor of our design studios. It turns out that this professor, Dr. Donald P. Greenberg was also the founder and director of Cornell’s nascent Program of Computer Graphics. What he showed us that evening utterly blew my mind and ultimately changed my life.
Eventually, Don invited me to join his lab as a graduate student where I became part of a group of pioneering graphics researchers who made seminal contributions to the development of interactive 3D imaging systems. It’s fair to say that the only art I’ve made since entering Don’s lab in the fall of 1981 has been natively digital. I’d have it no other way, especially now that after all these years this art form is finally getting the respect and recognition it deserves.
I should add that early in my undergraduate career, I took my first photography class and discovered a love for that medium. My education was also rounded out with numerous courses in architectural history, landscape architecture, urban design, and art history, all of which continue to inform my artwork.
BW: Can you tell me about the intent and process behind your Dreamscapes?
DA: The motivation behind my Dreamscapes project — and that’s how I see it: as one continuous art project that is now in its twelfth year — has always been driven by a strong desire to share as fully as possible my experiences of great landscapes and cityscapes.
It’s hard to explain, but I can be hiking all day in a beautiful place like, say, Glacier National Park, but every once in a while, something magical happens: the scene suddenly comes together in a particularly special way that sets alarm bells ringing in my head. I struggled for decades trying to capture and convey that experience with traditional photography and gradually learned that it takes much more than a camera to achieve that.
Dreamscapes are my solution to this challenge and are built on two key insights I had over the years. One insight (which I later learned David Hockney explored contemporaneously in his own way during his A Bigger Picture period) was that cameras see the world in a much more limited way than humans do. I needed to devise a computational photography technique that forces my camera to see the world the way we do: with a much wider field of view (both horizontally and vertically), with a much higher dynamic range (we have no problem seeing details in dark and bright parts of a scene), and with rich detail everywhere we look.
That breakthrough, which came to me in a flash in front of a canyon in Utah on September 27, 2011 (after years of incremental experimentation, of course), I consider the official start of my Dreamscapes art project.
The second insight was realizing that powerful landscape experiences are not only visual (seen) and visceral (felt), they are also cognitive; they make you think. At least that’s what happens to me when a scene is powerful enough to take my breath away; I start waxing philosophical about the nature of seeing and even the nature of reality itself.
When Google’s “DeepDream” came along in the summer of 2015, I saw the opportunity to bring a cognitive element to my landscapes and, with the generous support of two brilliant engineers, Joseph Smarr (Google) and Chris Lamb (NVIDIA), in early 2016 I was granted the “keys to the kingdom”: a modified super-scaled cloud-based version of DeepDream that could operate successfully on my giant images. I’ve been tapping the potential of this workflow ever since despite the fact that the original AI code base has not changed and was not even designed for making art in the first place! In fact, DeepDream was part of a computer vision system tasked with classifying images.
BW: Your Dreamscapes feel like psychedelic impressionism, where the guided zoom-in reveals emergent textures you wouldn’t expect. Who or what would you call out as your most prevalent artistic influences in creating this work?
DA: Again, the impetus for pushing my art in this direction goes back to my original motivation to create artworks that move people not just visually, but also viscerally and cognitively. That said, in the course of developing my Dreamscapes, I’ve become an ardent student of the history of landscape painting, and the work of three artists, in particular, resonates most strongly with me.
The first is Frederic Edwin Church, perhaps the greatest of the Hudson River School artists, whose immersive highly detailed master paintings like The Heart of the Andes drew crowds and created almost cinematic landscape experiences.
The second is Maxfield Parrish, who gave up a successful career as America’s most famous illustrator to dedicate himself to landscape painting in the 1930s. Parrish was a true role model for me; he tinkered with both old and new approaches to his medium and brought a level of vibrancy, detail, and luminosity to his paintings that I absolutely love and which hadn’t really been seen before.
The third is Eyvind Earle, the late great Disney background painter turned fine artist, whose stylized colorful landscapes verged on the mystical. Earle himself was somewhat of a mystic, as evidenced by his poetry, and he managed to imbue his paintings with a magic, mystery, and grace that is very inspiring to me.
Of course, there are many other amazing artists who have taken landscape art beyond pretty pictures and into experiential realms who also deeply inspire me: artists like Seurat, Monet, Van Gogh, Dali, and even Klimt.
Interestingly, DeepDream occasionally seems to channel all of these artists (sometimes in the same scene!) despite not being trained on their artworks or even having been designed to make art; DeepDream was a diagnostic tool developed to peek under the hood of Google’s image recognition AI; it is actively trying to make sense of what it’s seeing.
The fact that recognizable motifs of these artists seem to spontaneously arise from DeepDream’s inner workings indicates to me that there is something that this extraordinary AI and these visionary humans have in common the way their respective neural networks process and perceive the world.
BW: Your 3D work is really quite fascinating. Can you tell me why you created these? Do you see this kind of art expanding in the future as the technology for experiencing it becomes better and cheaper?
DA: My 3D work resulted from an idea I had about a unique way to showcase curated details of my Dreamscapes and, in the process, express my former life as an architect and 3D graphics researcher. My real interest these days in 3D revolves around the possibilities for exhibiting virtual sculpture in mixed reality.
After 40 years of waiting, I’m happy to report that I’m in my Quest 2 headset almost daily. Not only do I maintain a spacious social VR Dreamscapes art gallery where I’ve enjoyed both private and public openings with large groups of attendees, but I also meditate with eyes/ears open most mornings in either synthetic (CG-based) or realistic (photogrammetry-based) virtual environments.
I hope I live to see the day when I can immerse myself and others in sumptuous lucid dreams generated in real-time 3D using words or thoughts alone!
BW: As someone who seems to be constantly experimenting with different techniques, how would you characterize your creative process from a high level?
DA: I don’t experiment with different techniques for the sake of experimentation; at least throughout my Dreamscapes project, my experimentations have always been in service of my original artistic intent: to create grand scale luminous landscape art experiences that inspire my fellow humans.
As I mentioned previously, my quest has also motivated me to dig deeply into precedent and learn from past artists who seem to have struggled with and solved similar challenges. Even entire art movements of the past have inspired my creative process and led to experiments informed by Cubism and Abstract Expressionism, for example.
The neuroscientist, David Eagleman, maintains that human creativity involves “bending, blending, and/or breaking” what has come before. I agree with that; I don’t think anything is created in a vacuum. I don’t believe in original content, but I do believe in original vision.
BW: What do you think common criticisms of AI art get wrong?
DA: “AI-generated” art is a misnomer. AI is inert; it has no drive to express itself and certainly does not generate art of its own volition. The artist must have the idea, direct the AI, and curate its results. Technology has always been a part of the art-making process; more powerful technology expands the possibilities for artists, it is not subtractive. It can be somewhat disruptive initially but ultimately yields a net positive.
AI art tools that make it easier for people to realize their artistic visions are a very good thing in my opinion; personally when it comes to individual artworks, I don’t care about the time it took the artist to create, what tools they used, what training they have, how much they practiced their skills, or how much eye/hand coordination they were gifted with and/or developed over the years. If AI art tools could help an imaginative but quadriplegic individual fully realize original artistic visions, would that art have no value?
BW: What do you think common defenses of AI get wrong? Is there a middle ground between the two extremes?
DA: As I said, technology can be disruptive, and in fact, the latest AI art tools have already caused some economic disruption, particularly in the realms of art-for-hire (e.g., illustration, graphic design, concept design, storyboarding). These tools will indeed make it harder for people with traditional art skills to make a living and denying that to be the case is just wrong. Ironically, the people best qualified to make art for hire in a world of powerful prompt-based AI art tools will be those folks who double majored in literature and art history, who traditionally had a tough time finding work! They will have the best ideas and the ability to curate the best output.
Also, artists’ copyrighted works should not be automatically ingested from the web for training these tools and artists should not have to opt out of that; they should have to opt in. That said, I think artists would be fools not to opt in; art created in another artist’s style is fan art and will ultimately advance that artist’s position in the field. It’s easy to prove whose style came first, particularly if rules are enforced to expose “in the style of” metadata in artworks created with these new AI tools. I don’t think the original artist should be compensated for art generated in their style, but they should certainly be attributed if the derivative artist explicitly references the original artist’s work in their prompts.
BW: Do you have any tips for people who want to build a body of work and develop a personal style with AI tools?
DA: Start with an idea, have a motivation, an intent… and, most importantly, a clear original vision. And study. Read great works of literature. Visit every art museum you can. Attend art openings. And curate your work ruthlessly.
BW: How would you characterize your use of AI? Is it collaborative, or is AI just a tool, like a pencil or a paintbrush?
DA: AI is just a tool, but a tool like no other and at times it feels like a collaboration. The metaphor I like to use is this: I’m the leader of a jazz band who writes original compositions and I have a virtuoso saxophonist in my band who knows exactly where I’m going with my song, but who is going to improvise and bring their own sensibility to the piece. In doing so, they will constantly surprise and delight me. That’s exactly what happens with my AI tool/partner.
BW: How does frustration manifest in your creative practice, and what do you do to clear creative blocks?
DA: There are many places in my creative process where frustration can arise: uncooperative weather can certainly frustrate my photography work, software tools can become obsolete or break, shared computing resources can get overbooked, etc.
As for creative blocks, I don’t force the issue. I believe curiosity and creativity tend to cycle like the breath; after an extended period of creative output, when I’ve exhausted myself and can’t “exhale” any further, I let curiosity take over and I “inhale” content from others (e.g., art, books, movies, podcasts). At some point, that cycle starts over again. I just pay attention to the signs and act accordingly.
BW: What do you hope to accomplish with your art? What do you want viewers to take away from the experience?
DA: I remain mystified by the human experience of the world we live in, particularly by my own experience of these heartbreakingly beautiful places that I seek out and venture into. This is my muse; I’m not totally sure why. I think the universe is trying to tell me something profound about light and form and the interplay between the two, especially as witnessed in magnificent wide open spaces. And maybe part of what it’s trying to tell me is that it’s my job, with my peculiar mix of education, experience, skills, and talents, to deliver this message to my species.
What message? I can’t put my finger on it, but my heart tells me it has something to do with life, love, and consciousness. If my art does anything to move our species forward along these vectors, that would make the struggle of pursuing an artistic life worthwhile.