What characteristics define whether a news-photo is successful or not? During the past decade we’ve experienced an exponential growth in photography, images both taken and shown. This raises the question of how news editors choose images that will represent an event out of thousands of candidates. Are we as a society actually aware of the manner in which events and news gain a face? And what are the consequences of these developments?
We discussed these matters with an artist Coralie Vogelaar at her studio in a former school building in Amsterdam. For her beautiful new work Recognized / Not Recognized, Vogelaar scraped the image databases of international news agencies such as AFP and Reuters for images of the 10 most covered events of the past five years, amassing a database of over 850,000 images. Using image recognition software she attempted to find underlying patterns, as well as tried to determine which images were used most often in the visible part of the web indexed by Google. Is it possible to determine emotional or aesthetic preferences in the images selected for publication?
Vogelaar told us about her interest in analysing enormous quantities of images using contemporary technology. “Google provides us with the opportunity to show how often an image has appeared on the Internet, which I believe is a fitting way to measure success factor of an image”.
She decided to focus mainly on witness photography in the project. “I find interesting the political influence of these types of imagery”. However, she has consciously avoided selecting the top 5 most iconic photos of these events, as these have already been analysed and criticised thoroughly. “Of course I could have selected an image of Alan Kurdi, the boy whose photo went viral after drowning in the Mediterranean Sea. But I wanted to avoid realising work that is about one specific current affair, that only draws attention to itself. For this reason, the question is too comprehensive.” In her research she focused more on one layer underneath; making visible the underlying mechanism by which images become part of our news narratives.
Her final collection lent itself well to a cross-sectional study. The core presentation of this is done using movement study. These results were interpreted by choreographer Marjolein Vogels, who designed choreography for nine dancers, in order to represent both the ‘successful’ and ‘unsuccessful’ news images.
The result is an intriguing video-installation consisting of two screens. The right screen shows the successful news photography, and the left screen shows the unsuccessful news photography, taken on almost the same moment and at the same spot. Both beautifully and carefully translated into choreography. “The poses of the dancers are abstractions of the compositions in the news photograph,” Vogelaar adds. Her choice for dance is well-considered: “I’ve decided to work with a choreographer, as choreography is concerned with posture, characters and framing. These elements form the basis for my analysis of news photography.”
The video installation is supported by the rich series of the successful and unsuccessful news photos, where the composition study is based on. This photo database gives insight in an almost machine-like way of analysing lines and faces that grab attention in an image. It is a search for the ‘successful photo algorithm’.
The exhibition opens with a collection of ‘orphaned’ images from a single day of a single event; photographs that are impossible to find on the open Internet, and as such are considered redundant data. They only exist behind the login pages of the databases of the press agencies.
Vogelaar assumes a scenario in which the management of large-scale data sets will be done in time by computers. “Now I can still see unsuccessful photos behind a login but In the future, unsuccessful photos will be hardly findable, simply because they’ve been filtered by algorithms and other techniques inside the camera or database”, Vogelaar posits. Especially the way in which images are objectified she finds interesting. After all, the algorithms do not come into being on their own. In there construction certain assumptions are made regarding the input that these algorithms will receive in the future. Choices made at this stage will have consequences for the manner in which future events and news facts will gain a face.
As such Vogelaar raises an urgent debate using Recognized / Not Recognized. Who decides the characteristics in this selection? In Vogelaar’s opinion both audience and creators (so society as a whole) are responsible. By introducing software selection systems, a seemingly ‘objective’ third agency is introduced into this equation. What will it mean for the quality of journalism if only objectively 'successful' images reach the audience?
However, there are only a few people who notice the importance of computer-controlled selection of news photography. Another question raised is what these types of services will bring to the table. What if a news agency offers to ‘optimise’ photos based on the attention an article is getting from a certain target group? In these cases one can only speak of a computer-controlled editing process, in which the editorial character of a picture editor is side lined. This way, news photography can become brittle and vulnerable. Only a few channels will have to malfunction, and the whole of journalism falters, or behaves in a way that is different from the desired behavior. In that sense, it is important that we depart from the idea of an algorithm or system as an objectifying measure. However, computer-controlled we may become with every one of us, we remain dependent on human situations and values.
In Recognized / Not Recognized Vogelaar has her feet firmly planted on the ground, because she leaves the debate around her work to the audience. We tent to choose from the enormous sea of photographic data, and therefore find those images that suit our eye and affirm the Western compositional tradition we are all surrounded by. Google’s algorithm is already built on the validation of certain pre-sets and stereotypes. At the moment — with a little bit of effort — unsuccessful photos can still be found behind a login wall. However, in the future these images would not even be registered by the camera, and if the camera delivers them than it is very unlikely they can be retrieved. Unsuccessful news images are then being considered as dirty data and are simply wiped from the system. By analysing and deconstructing which preferences we seem to have regarding news photography, already much ground will have been won according to her. We are curious to see where her developments lead.