Intro
What is the use of a photograph? If nothing else, it’s a reference point to a specific moment in time, the moment of its creation and the moment it depicts. It is also about the subject it depicts, the occasion, the person, the act, or the scene it shows.
Especially family photographs are a tie to our personal history, an archive of whom we are and where we came from. They are “memory deposits” from where we keep sourcing the certainty of our life events even if our own memories of them are lost.
The photographer, in this case, is usually someone we know, our dad or mum, aunts and uncles, family friends and relatives, someone we would invite to be next to us in the most important events of our lives, someone we would open our house to, someone we would trust with a funny face, an embarrassing outfit, a personal moment. A human.
Though, it’s a long time now since humans are not the only creators of images anymore. Automated cameras, various scanners, Artificial Intelligence systems, and GANs are all able to produce images.
In this new project personal memories are absorbed by A.I. and meshed with thousands of others to become a new breed of images, images resembling in an uncanny way those of our childhood photo albums although everything is translated to a mere set of forms, points, colors, and tonalities.
But what is the use of this new kind of image since it loses its fundamental tie to reality? What is the memory they hold? To whom are they precious, heirlooms that must be preserved?
Plato would probably classify them as third-grade mimesis and thus reject them as a means of even further diversion from the truth of Ideas, but is there any chance that Aristotle would appreciate them for the insight they may give to the learning process and the potential of catharsis that they may offer?
If the danger of photography was our belief in its inherent truthfulness (a belief that has been challenged ever since) may there be the same danger in qualifying AI-generated images fake? Fake in comparison to what? Untruthful to what kind of truth? AI-generated images are derivatives of photographs, an amalgam, an average, of billions of other images, yet may that bring them closer to a more universal kind of truth of phenomena?
The etymology of the word image has its roots in the Latin word imitari, meaning "to copy or imitate"; and the words imagine, imagination, and imaginary derive also from the same root. Although today we use the word imagination usually in a context that predisposes us to something new, that newness seems to derive from the ability to imitate and copy the already existing. In that light, AI-generated images are ontologically equal images to any other image preceding them.
AI-generated images have a peculiar relationship with time. Unlike photography, they are not tied to a specific point in time through the scene they depict, rather they can imitate this link. Trained on billions of images, AI has assorted the style and color grading of the medium humans have used to create images in the past. Especially when a photographic result is requested specifying the time and the place that an image was created, for example in the 60s, AI adjusts the colors and the overall feeling of the generated image accordingly. But how will we be able to date AI images when looking back in time? Deformations, distortions, ambiguity in the form, and the “in-house” style are the giveaways.
Moreover, AI-generated images have no actual form of their own. They tend to mimic the style and aesthetic of other mediums, be it painting, photography, or digital art.
2023, Installation view, Photo Vogue Festival, Uncanny Atlas: Image in the age of A.I., BASE, Milan, Italy
Imagined Images (in progress)
My great-grandparents, my grandparents, and my parents changed the place of residency many times before I was born. Sometimes forcibly, sometimes at will. The photographs that documented the events of their lives were lost along the way. I don’t really know much about them, except for some stories I’ve been told about where they lived and what their profession was. I took all these stories and used them as prompts in an image-generating AI to rewrite my own history. Unexpectedly, that process was not only emotional but also informative. The AI seemed to know more than I did about a specific place and time, adding details to images that I wasn’t aware of.
Moments that happened, moments that were unphotographed, moments I imagined, moments I was told about, moments I have hoped to happen, moments that never happened.
How can one rewrite his own history?
_______________________________________________________________________________________________
The project “Imagined Images” focuses exactly on the intersection between the most intimate human photography and AI-generated images while also highlighting the correlation between words and images. It reflects on the popular, amateur, use of photography to document family history and how our perception of self and belonging is formed through those images. The project also touches on the topic of curating practice, since the family photo album is highly curated and therefore a constructed history of the family. Noticing the similarities of the photographs included in family albums made me realize that photographers follow, unconsciously, an unspoken script of what a family album should include. This implied scenario of a normal and happy life is transferred between space and time through images and not through words but can be so imperative that if someone is missing parts of this fixed story he may even feel deficient or inadequate. (For example, I don’t have a single photo of a birthday party in my album, because I never had a party for my birthday, so I felt I needed to create many of them with AI. In this way I have added many images that never actually existed, to fill the gaps in my family history and recreate it in a way that would help me reconcile with my actual story. ) Through the use of image-generating AI this scenario becomes visible again and is used as the text prompt that creates the image. The fact that the faces in the images are unrecognizable due to AI’s early version’s inability for photorealistic results seems to intensify our ability to relate to those uncanny images. They are not mine, yours, or anyone’s, but they belong to all of us. The aura is there. The surface of the image is just a stimulus, It’s a certain point that functions as an anchor to a place and a time, It’s a directory, a path, a link to memories and feelings. We used to say that “an image is worth a thousand words” but today a word can produce infinite image variations, reversing that relationship.
