Opinion: From Awe to Oct. 7, a San Diego Digital Artist Explores the Meaning of Israel in Imagery
As a photographer of Israel, I've thought long and hard about what my images convey. Now with AI, even when the generated image is wrongheaded, it is a way to ask myself, "What is Israel?"
There must be a billion photographs about Israel from the birth of photography to the present.
I wondered how I could capture the sense of the place and cultures within it. How could I find a different and intriguing approach? I looked back to ethnographic research in Cali, Colombia and in projects in the U.S.-Mexico border region. I could buy a more expensive camera. I could also create my own art. Here is what I learned to see over nearly forty years.
When I first traveled to Israel in 1985, I was a tourist. Many of my wife’s relatives lived there, having escaped or been forced to leave Iraq in the early 1950s. That was when I snapped photographs. I had yet to become a proficient digital artist.
It wasn’t until the late 1990s that I would be able to repurpose my earlier photographs. An adventure in visual art and my relation to Israel began to emerge. I began to impose meaning on my travels to Israel. It was not just being there, but being there in multiple ways with the changing technology of photography and the digital darkroom.
My photographs are often edited as digital art. In one photo from 1985, I deleted a view of the Kidron valley below the way up to the Temple Mount. There was a line of rabbis overlooking a low wall. I saw the valley as a distraction. I added clouds from my home in Poway. I now found a sense of awe in the image.
I don’t know whether the rabbis had experienced what I imagined. But my image pictured it that way. Years later, I was surprised to see my image resonating with Exodus 24:10 — and they saw the God of Israel — under whose feet was the likeness of a pavement of sapphire. Perhaps I was now seeing what others had imagined in biblical times. In one way it became my just-so story.
Places of Significance
When I revisited Israel in 2007, I became intrigued with digital infrared photography, The structure of the photographic image remains the same, but the color and lighting sensibility changes. Israel became infrared.
This was about using photography, and editing it, to pay attention — a way of mindfulness. Henry Miller captures what I was finding in Israel in blades of grass, about this sense of paying attention:
“The moment one gives close attention to any thing, even a blade of grass it becomes a mysterious, awesome, indescribably magnificent world in itself.”
So, too, I was finding the places and cultures in Israel. In the moment of taking photographs and inquiring into them with image editing.
Israel in a Smartphone
I was among those trading in their cameras for smartphones. Perhaps the loss in high resolution was not a loss at all. The photos could retain their structure while sacrificing a bit of sharp lines, texture and color.
For most photos, the loss was barely noticeable and whatever was lost could be mostly regained with digital editing. Perfection continued to give way to good enough. But the technology was not static, it improved.
It was now 2023 and I decided to rely on my smartphone camera. The important question for me was weighted toward the art and less the measures of technical accuracy.
Could AI Model What My Camera Saw?
Once more I went to Israel. COVID had delayed my travels. It was now 2023. I began to experiment with generative AI to create images. The most accessible platform was DALLE2 (an OpenAI technology). My plan was to take photos with my smartphone; then I would type words that described what I was seeing into the generative AI platform. My plan was to see if my words could capture what I saw. These two images — the actual photo and the AI replicant — would frame a conversation for me.
How do these two realities gibe with each other? One seen and the other energized by my words and filtered through the layers, loops and algorithms that had a database of millions of images on which it had been trained. This was an experiment in “seeing” or understanding what it is that we tell ourselves what it is we are seeing.
Often, I would have to change my wording to reach a more accurate approximation of what I was seeing. I admit this was a strange experiment. But it was an intellectual exercise that challenged how my words reflected what I perceived. What I was rewarded with was not what my photograph pictured so much as a surprise.
Even when I found that the generative AI image was wrongheaded, it was a way to ask myself, What is Israel? The way these generative AI models crafted its versions of my text prompts was biased and puzzling, then and even now.
I found myself in one moment having a coffee and pastry at a local café. As my wife and friends chatted, a woman got up. She was juggling her coffee, phone and cigarette. I tried to capture that moment with this text prompt: photorealistic image, middle aged woman, just finished her coffee outside outdoor cafe, cigarette in one hand, cell phone and coffee in her other hand.
The original photo on the left is an example of street photography. The woman, outside a Tel Aviv café, is struggling to balance her coffee and cell phone (and possible her purse as well) while smoking a cigarette. The images to the right are examples of using a description of the woman as a prompt in two different generative AI models. The DALLE2 image lacks the near perfect rendition found in Epic Realism. Neither, though, capture the entire detail provided in the text prompt.
As I tweaked the test, the generative AI models continued to improve. But an important question remained. In using AI, there is no human authorship unless one considers the text prompt sufficient to qualify as the human element.
Still, there is potential benefit. If one could not access an unfolding reality, perhaps the AI platform could help imagine it. Picasso had imagined the horror of Guernica. Could AI do as well for the unfolding horror of Oct. 7?
That was what wondered as we waited in our Tel Aviv apartment. Our thoughts tried to catch up to the unfolding horror.
Imagining Oct. 7
I tried to imagine what was happening in our building. We were shown the stairs outside our apartment. There were already other residents sitting there. My text prompt to DALLE2 described an odd staircase with individuals oddly grabbing each other.
The generative AI images often fail to picture the reality we can see; it may refuse to honor our words because of its algorithmic biases, company policies and guidelines; it may be unable to translate our words into what we perceive. In this sense, AI images fail to supplant our own creativity.
I was determined to capture what I experienced on Oct. 7. Since coming back from Israel, I started working with other generative AI platforms including Stable Diffusion. This allowed me to keep the same structure of the image while adding a multitude of novel and unexpected elements.
I returned to one of the photos I took on Oct. 7. We were shown our building’s bomb shelter by one of the other tenants. It was a basement room that appeared never to have been used. The photograph pictured a drab stairwell leading down to the building’s safe room. I wanted to bring that stairwell into the aesthetic that drove Picasso’s Guernica to Dali’s Premonition of Civil War to Goya’s The Third of May 1808. This stairwell was not the horror of the south of Israel; it was a different fear — a tourist in Tel Aviv.
I would be naïve to think that picturing reality would be better with a camera, AI or digital image editing. Or a paintbrush. Or with any tools an artist might choose. Our eyes center reality, but even then we are deceived. Still, we try to understand the reality in which we exist.
Joe Nalven is a San Diego-based digital artist. He is the author of Going Digital: The Practice and Vision of Digital Artists (Thompson, 2005). He will be giving a public lecture at the Coronado Public Library on Jan. 15 at 10 a.m. The lecture is free and sponsored by the Center for Jewish Culture. For more information: Tel. 858-362-1150 or www.lfjcc.org.
What's Your Reaction?