The title of this post was generated by AI, the content is my own thoughts.
“Generative-AI tools will not wipe our memories or corrupt our sense of self, but the camera roll of the future could be missing the little bits of texture that allow us to recall certain moments as they really happened. So often, technology promises to decrease the friction in everyday life and to optimize how we present ourselves to the world. In the moment, a flawless AI photo is seductive, but the appeal can be lost when you’re looking at it years later. A perfect camera roll is not one that’s full of perfect photos.”
The image above is from Sasha Reiko Photography
This quote reminds me of another one in a book called Camera Lucida, by Roland Barthes. He speaks of his mother’s death and trying to find a photo of her that genuinely felt like her- one that completely captured her likeness, but each one always fell short of the task.
“Photography thereby compelled me to perform a painful labor; straining toward the essence of her identity, I was struggling among images partially true, and therefore totally false. To say, confronted with a certain photograph, ‘That’s almost the way she was!’ was more distressing than to say, confronted with another, ‘That’s not the way she was at all.’”
We think of photographs as showing the truth, but I think all of us can relate to the feeling of taking a photo and then looking at it and realizing that the recording of what happened will never be truly as sensational as the actual experience is, or that the image just can’t replicate what you can see with your own eyes. Cameras shorten or deepen a field of view, expand or compress space, change the color balance, over or underexpose an image, and more. Cameras can get close to capturing what a human eye sees, but it cannot. The whole truth of our experience cannot be present in a photograph, as much as we’d like to think that it can. This of course does not mean that the truth cannot be informed by photographs, but a photo by itself is not the truth.
Adding generative AI to the mix creates a whole new level to the subtle distortions cameras already have within them. In Camera Lucida, Barthes goes into a long discussion about finding a singular photo of his mother when she was 5, which to him satisfied his grief-stricken longing for “a just image, an image which would be both justice and accuracy–justesse: just an image, but a just image.” He saw “the distinctness of her face, the naive attitude of her hands, the place she had docilely taken without either showing or hiding herself, and finally her expression” and he felt like he finally saw her for the first time since she had passed. Imagine if, instead of living in the early 1900s, Barthes’ mother grew up in the 2020s. This single photo which meant so much to him might have instead been adjusted in a way that made it “partially true, and therefore totally false” just like the rest of them. The authenticity of the moment captured would be lost and he would not have been able to feel that connection to his mother that he was searching for so desperately.
Flawless AI photos cannot replace memory or its next best replacement, photos taken in earnest. They entirely miss the point of why the photo was taken in the first place. Often what is most interesting and important about a photo is what is unique or even ugly because that is real or at least genuine. A perfect camera roll is one that has heart and intention behind it- not something that Artificial Intelligence is likely to develop any time soon.
Leave a Reply