Algorithms & Automatic Curation - An Interview with: Adam Griffiths


Forme Journal is a curatorial project that occupies the space between fashion and fine art. It’s first manifestation is a print publication, within which its’ designer Adam Griffiths applied an algorithm to create an automatic curation process, creating disparate relationships and arbitrary compositions between the images. We caught up with Adam, a Design lecturer at Manchester School of Art, to learn a little more about this process and its’ implications within curatorial practice.

Firstly, can you tell us a little about the algorithm – what does it function around? Does it adhere to any set visual parameters?

My whole fascination with generative processes and particularly the algorithm in question started completely by accident. A happy accident of sorts. It stems from a lot of research I was doing in terms of our very contemporary, millennial attitudes towards collecting online imagery, be it through Tumblr, Pinterest or generally just saving and hoarding into folders and files.

I discovered the algorithmic process between saving and exporting stacks of images between certain Tumblr themes and the computer’s print settings. I’m currently in the process of formalising this work in someway and so I’m having to keep some of my tricks up my sleeve but in a nutshell – the algorithm occurs when trying to translate online/screen content to the printed page. It becomes a fascinating collaboration of sorts between myself, the web and the printed artefact.


The process adheres to a number of basic visual parameters but beyond these, the results are completely arbitrary. Firstly, I’m very much in control of the content that is fed to the process, be it my own photographs or found imagery. I also have to control the output for the process, the size of the printed page(s) for example. But, in terms of the composition, ordering of sequences and scale of the content, these are in the hands of the system itself.
The translation from screen to the printed page creates a number of key characteristics that I’ve adopted and established within my practice. It’s an almost anti-design, anti-curation system in which the algorithm breaks all the concrete expectations that we often think makes good curatorial decisions or clear visual communication. The system becomes the curatorial anarchist, resulting in randomised page compositions, overlapping of disparate photographs, extreme cropping and non-linear sequences. The algorithm is essentially a tool for disruption and fragmentation in multiple fashions.

Applying the process to a linear or strict series of images, conveying a narrative for example, results in very fragmented compositions in which the linearity becomes blurred or completely distorted and therefore the reading of these original contexts become distorted. I’ve quickly realised that it’s this notion of disrupting that I’m truly drawn to.


Can automatic curation justify its’ decisions like a human might be expected to, or would that defeat the object?

I think the most exciting and often successful element of automatic curation and publishing, is that it requires the human decision to take a number of steps back. In my case, I’m letting it take over the steering wheel once I’ve supplied the process with the visual content. The most curious element of this approach is that I’m in no way in control of the output from a design aspect, which for me is truly liberating.

I’m sure other and more sophisticated modes of automatic curation could justify their decisions – but more than likely that would require the touch or assistance of the human programmer or editor which in turn would essentially defeat the purpose. If we want automatic curation to offer us something different to the humanist kind, then we truly need to stay true to its synthetic behaviour.

In regards to my process, I still don’t understand any of the technical sides to what I’m implementing, or why/how it does what it does. I think that’s what has made it successful for me in terms of underpinning a lot of my work. The minute I understand or can control the process is when it becomes defunct and is presented as my directorial vision, not the unsanctioned vision of the algorithm itself.


Forme’s opening statement notes that, as a curatorial project, it will take influence from its surroundings to determine and shape its evolution. How do you envision the evolution of broader curatorial practice?

 I think we live in an exciting time creatively and conceptually, by which I mean we can create inbetween parameters and spaces and have no strict boundaries of restriction. Something I was really excited about when working with Forme and founder/curator Lucie Crewdson, was the prospect of Forme not being bound by format. It is a curatorial, explorative project that focuses on the space between fashion and fine art photography, but that doesn’t conjure a specific format to explore it within. Issue 01 took the form of a printed publication, Issue 02 may be an event or even an object, all of which would require a distinct direction and sense of curatorial value. We live in such a blurred space of cross disciplinary practices that I believe organically these broaden and impend on new territories, and so curatorial practice is very much apart of that.

I believe that one of the main roles of curation as a practice is to build bridges within and beyond our comprehensions of modern and historic culture. How might the development of automatic curation influence the future role(s) of the curator?

I feel like the process of automatic curation allows us to obtain very unexpected results that go beyond our classic idea of what ‘curation’ is or should be. This method may influence the future roles of the curator dramatically, or could completely be ignored. But it’s an added dialogue that can be thrown into the mix of possibilities.
I’m interested in how we may embrace the curatorial vision of the machine to present works much as we embrace the machine as a tool to often produce it.


If the algorithm’s activity were applied to a physical gallery space, how might this compare to more traditional curatorial models?

This is something that I’m truly intrigued by at the minute. A distinct characteristic of the algorithm is to break up linear sequences, and to re-present them in a much more dynamic sequence of scale and composition that truly fragments the original contexts of the works. Applying this to a physical space could be easily translated, the process could determine not only the order of works, but perhaps positioning within the space, overlapping of disparate works or artists, creating new discourse around works, practitioners and concepts. To put it bluntly, if we were to place Andy Warhol’s ‘Electric Chair’ (1964) image over Benvenuto di Giovanni’s ‘The Crucifixion’ (1491) would we read both images separately, or would we begin to read these simultaneously as one piece? And what may come out of that?
I’m currently in the process of adopting the process in a much more physical setting, as a form of ‘print performance’ in which the process becomes the work, aswell as the visual output.


What might we learn from further exploring and experimenting with this process?

Marcel Duchamp coined the phrase ’Infrathin’ to recognise some of the indescribable spaces or occurrences between states. He described these in examples such as the warmth of a seat (which has just been left) or when the tobacco smoke smells also of the mouth which exhales it, the two odours are married by ‘infrathin’.
I see this in relation to the algorithmic process, further investigation may give us more clarity to swim between the online and physical states which for me, could be fascinating.

Thanks, Adam

Website / Instagram

Forme Journal