Today, I took part in an interesting panel conversation at Fotografiens Hus in Oslo in connection with the exhibition What We Call Real by Camila Urrego. This was the first time they had invited a photographer to exhibit works explicitly made “with AI,” and I was invited to contribute thoughts on the potential of AI for creative practice (in this case, photography). As MishMash picks up speed, I receive many requests for talks and panels on AI and creativity, and I try to say yes when I can make it. I have found this to be an excellent way to think about various AI use cases outside my core focus areas and to meet interesting people. Here are some thoughts based on my preparatory notes and reflections after the event.
My own use of AI in photography
I am a hobby photographer, but have never really worked with AI-based tools in photography. However, since many of my photos these days are taken with my mobile phone, I, like most others, rely heavily on backend AI processing when using the “auto” mode. When I am thinking more photographically, and not only photographing for pure documentation, I often turn on the “pro” mode on my mobile (currently a Samsung Galaxy Ultra S26), which appears to be more “manual” in its settings and uses less visible AI magic by default.
The “objectivity” of the photography
As far as I can see, there have historically been two “camps” in photography. One is the group that sees a photo as documentary evidence, and in this case, the image should aim to reveal the world as it is.
However, I (and many others) would argue that photography has never been objective. When you frame something in a photo, you choose to depict that thing rather than something else. That choice is already “removing” a large part of the scene.
While “photoshoping” has enabled digital photo modification over the last few decades, altering images has been possible since the beginning of photography. I still remember working in the darkroom at high school in the 1990s, where we learned how to remove objects from a photo by gently covering them while lighting the photo paper. Today’s AI-based removal algorithms are just continuing the same.
The exhibition
Camila Urrego’s What We Call Real served as an interesting backdrop for the conversation. We never got to hear details about what she has been doing, except that she has photographed herself in nature, has “used AI” to work with the photos and has then made a selection that has been printed on paper to be displayed in the gallery.
It was fascinating to see how well the final photos—if they can still be called that after all the manipulation—worked in blending the human-made and the machine-made. Some were clearly more machine-like than others, but all showed some level of humanness in portraying Camila in different landscapes.
One question I had for Camila was whether she thought of AI as a “tool” or a “co-creator”. Interestingly, she argued that it was both. At times, a tool, other times, a co-creator. This is probably how made creatives explore AI these days, trying to make sense of what works and what does not. In this dynamic, the latent space becomes a space of exploration, not just a technical one.
The “Fourth Narcissistic Wound”
Also on the panel was psychologist Gunnar Gjermundsen, who framed AI as humanity’s “fourth narcissistic wound”. The “first wound” came when Copernicus discovered that the Earth (and humans) are not the centre of the universe. Second, Darwin found that we are not as unique as we like to think, but part of the animal kingdom. Third, Freud argued that we are not masters of our own house. Now, AI challenges our identity as uniquely “thinking” and “intelligent” creatures.
The existential question is: If a machine can think and create, what is left for humans? Gjermundsen pointed to consciousness and sentience as the final frontier. A machine may imitate a feeling, but it has never felt anything and will never do so.
The Materiality of AI
The panel was led by art historian and journalist Silje Sigurdsen, founder of Kunzt.no. She challenged us to also reflect on the materiality of AI, highlighting the massive power consumption of data centres and processing.
My response is that we’re focused on creating small language models that run on people’s phones and laptops, including older hardware. Interestingly, my experience is that small models are much more engaging for creatives.
Lived Experience
To summarise, we all agreed that there is a place for humans in the future. We are at a point where AI-based systems can create “photos” that are indistinguishable from human-taken photos. So they have far surpassed the Turing test.
The same is true in music, and soon in film as well. The question then isn’t whether it is possible to replace humans with machines; it already is. However, is it interesting?
While AI can master the how of an image, the why remains in the human domain, while many are curious about new AI-based tools, I also see an interesting “back to basics” movement among my students. They collect LPs, record on analogue tape, and buy film cameras at flea markets.
Ultimately, art is a meeting point. Whether the tool is a lens or an algorithm, it is the human heart beating behind it that gives the work its value.
Thanks to CoPilot for beginning this blog post based on two pages of handwritten notes.