THE CULLING OF PHOTOGRAPHIC IMAGES

A cull of images is coming, an automated decimation, which will transform the ontology of the image once again. Google announced that from June 1, 2021, they will stop the free back up of compressed “high” quality photos. The significance of this development is hard to overestimate. At this point, any connection of smartphone photographs to the original medium of photography and associated practices will finally be fully severed. In the recent decade, substantial discussions about networked images accounted for some of the core principles of image operations today1:

A cull of images is coming, an automated decimation, which will transform the ontology of the image once again. Google announced that from June 1, 2021, they will stop the free back up of compressed "high" quality photos. The significance of this development is hard to overestimate. At this point, any connection of smartphone photographs to the original medium of photography and associated practices will finally be fully severed.
In the recent decade, substantial discussions about networked images accounted for some of the core principles of image operations today 1 : people took over a trillion photos in 2020, an amount that cannot possibly be processed by the human eye; images changed their nature from being a cherished photograph or, in any case, an object bounded by the limitations of its medium, to a digital stream of imagery, co-produced by the technical apparatus and circulating in the networks; the material of images is a continuum formed between photographs, 3D models, animations and environments, maps, layers of augmentation, and virtual production; images freed from sentimental, artistic, or instrumental meaning form databases used to train machine learning models to recognise and label images and derive conclusions and predictions about humans.
Nevertheless, services still print photographs, shops sell frames, and the media of desire, from design magazines to film, regularly feature photographs on the walls. The photograph in your phone, machine-assisted and -read, circulating in the network and readily informing data analysts on its host, was still-up until now-a photograph of a printable quality, lying in wait for a possible welcome into the paper-based world. The possibility to print, even if rarely used, still grounded image-taking in the originary photographic processes, whose ontological concern was the THE CULLING OF PHOTOGRAPHIC IMAGES

Olga Goriunova
The Nordic Journal of Aesthetics, No. 61-62 (2021), pp. 94-97 manufacturing of a material impression of the world. This faint connection, ontologically troubled by the materiality of digital media (electric impulses, Boolean algebra, discreteness), will now be completely removed.
When free backup is turned off, and memory cards fill up quickly, who will be able to afford life-long monthly payments for storage?
What will happen to family photography? Will the relatives of the deceased keep paying for the family archives, growing more and more enormous with every year? A possibility to lose a life's-worth of photographs will arise. People will need to assess the value of every image they take and decide which ones to save. Although Google or Apple will help us decide, deleting images of "poor quality," our style of taking photographs perpetually, formed by the availability of smartphone and network socialisation, is unlikely to change. Amazon and undoubtedly many others will provide cloud backup services, but an uninterrupted perpetual backup is a fantasy. We will have to accept that photographic images are no longer printable and their existence is confined by the strict bounds of resolution. The overflow of smartphone images into art exhibitions will stop. "Amateur" smartphone photography, despite powerful cameras and intelligent software, will basically be confined to the status of thumbnails.
The mass culling of photographic images can only mean one thing: the automation of image analysis does not require printable quality. Machine vision is modelled on human vision and there is no detectable interest in trying to liberate AI from the diktat of its limits and to explore what a non-human, truly computer vision could see. 2 It is the images made by humans that form training datasets, whose function is to prepare computational models to analyse more images made by humans. Such automation of imageprocessing is entirely human-centred because it aims to derive actionable insight on the human as the "source of all possible cash".
Clearly, unlimited data storage on the cloud is not ecologically sustainable. The concerns of preservationists of media art, net and software art archivists, battling with defunct software and hardware will now be everyone's preserve. Perhaps, photographic studios and family photography will come back. Perhaps, imagemaking of a classical type will be resurrected alongside the increasingly ephemeral streams of images, pixelated shadows, which will still carry the substantial information for AI, marketing, policing, the management of the population and individual The Culling of Photographic Images 96 consumers. Similarly to how facial biometrics is most concerned with the geometry of the cheekbones, nose and eyebrows, which form the "human face" for the assessing program, schematic images can emerge as a new genre. A dream of technocrats, keen to sort information into the categories of primary and secondary (and so on) importance, such future digital imagery of reduced resolution will carry minimal visual excess: just a little bit to titillate the social networkers and enough to meet a machine learning model's specification.
OLGA GORIUNOVA is Professor at Royal Holloway University of London and author of Art Platforms

NOTES
The Culling of Photographic Images