Editorial: Curating, Biennials, and Artificial Intelligence

This volume of the Liverpool Biennial journal Stages draws connection between Artificial Intelligence (AI) and curating, at the time of the 11th edition of Liverpool Biennial The Stomach and The Port, and against the backdrop of the global pandemic, political and social turmoils, and technologically mediated and sustained world at present. [1]

Considering the rapid developments in automation (such as AI) and how our relation to it has changed, it poses questions about the implications for contemporary art;  the limits of and possibilities for curatorial practice under these conditions, and the relevance and future of cultural institutions and global biennials in particular in the post-pandemic world. What are the lessons to be learnt. What can the practice of curating learn from AI, what can AI learn from curating, and how can both learn from questioning knowledge forms derived from colonialist frameworks of humans and machines?

Rather than a theme, The Stomach and the Port explores the body, drawing upon non-Western ways of thinking and knowledge production. The artists and thinkers gathered in this edition of Liverpool Biennial challenge an understanding of the individual as an autonomous, self-sufficient entity. The body is instead seen as a fluid organism co-dependent on others, continuously shaped by, and shaping, its environment. When our answers are drawn from a foundation of knowledge steeped historically in Western reason and frames of thought, a social understanding of what constitutes the human has assumed a particular singular body: that of white man. Women, LGBTQIA, black and people of colour, indigenous people and nature, are located in a space of lacking, in a place of disadvantage as well as subordination. Therefore, borders are not only geographic, but political and subjective, an outcome of historical processes created by the constitution of the modern/colonial world.

In the West, the brain has been designated the commander of intelligence. While our bodies inhabit the world, the brain processes our experiences and transforms them into knowledge — this knowledge then informs our understanding of the capacity to process the world. But knowledge is not randomly produced, nor legitimized, as definitions and forms of classification control the production of knowledge, and therefore the formation and reformation of subjectivity. How we can re-calibrate our sensibilities and include a plurality of intellectualities — not only coming from the brain — and to diversify knowledges of the world? Can we bring bodily organizational force of experiences, feelings, knowledge, environments and technologies together?

A parallel problem runs through a history of artificial intelligence, where the brain (or mind) has been a predominant metaphor, similarly steeped in instrumentalised notions of Western rationality and reason. At the same time, it might be possible to begin to think outside of these models and to look for other frameworks that not only include indigenous knowledge but non-human knowledge. This is not a naive position - machine intelligence is fraught with problems, not least how the models tend to replicate already existing gendered and racial biases, and established hierarchies and structures of power. However there are also ways out of this thinking, once we can understand and articulate these social and technical frameworks sufficiently well to be able to reconfigure them otherwise.

These are active debates in critical AI[2], and the ones which provide a means through which to not only reflect on parallel issues inherent to the contemporary globalised art world — and curating — but to go beyond existing paradigms. What kind of future infrastructures and curatorial practices can develop from the coming together of diverse human and non-human? What new modes of expression and vocabularies are possible? What new understandings, entities, relationships, and practices can emerge through the exercise of biennial making once open to the possibilities afforded by expanded human and machine epistemologies?

Reflecting these ideas, the title of this volume refers to a short text / research proposal ‘The Next Biennial Should be Curated by a Machine - A Research Proposition’ included in this volume.[3] Other contributions include existing writing and projects by Nora Khan, Suzanne Treister, Elvia Vasconcelos, Kate Crawford and Trevor Paglen, Victoria Ivanova and Ben Vickers, alongside new contributions by Murad Khan, Eva Jäger, Leonardo Impett, Magdalena Tyzlik-Carver, together framing these discussions across diverse fields. Underpinning the discussion is a Glossary — an extract derived from Winnie Soon’s and Geoff Cox’s book Aesthetic Programming (2021) — to provide a shared vocabulary for this volume.

The various contributions not only question the forms through which we formulate these discussions today, but point to new possible directions. In her essay 'Towards a Poetics of Artificial Superintelligence’, Nora N. Khan calls for new language, new imaginaries beyond anthropomorphism, ‘to access what we can intuit is coming but can’t prove or describe directly’; metaphors that ‘bridge the human and the unknown’ and that can ‘help bridge inequities in rate and scale’. As her title suggestes, there is a future world emerging in which humans are not the central intelligence but ‘irrelevant bystanders’ to Artificial Superintelligence. What possible forms this might take is explored by artist Suzanne Treister in her 2018 project MI3 (Machine Intelligence). It uses Google’s Machine Intelligence (machine learning algorithms) to create and process bodies of datasets to eventually result in new works of art,  presented for copyright free download and print. These new works are 'images containing the original source data of their own making, ghosts of the 3 created Machine Intelligences transmuted into the style of a dead luminary artist, visions which may travel into the future, inserting themselves into homes and spaces across the globe, witnesses, for an unascertainable time span, of whatever is to come.' The process is visualised in a diagram presented alongside description, images, and notes.Taking a similar diagrammatic strategy, Elvia Vasconcelos’s contribution A Visual Introduction to AI, presents a collection of sketches intended as accessible maps to the history of AI and the basic components of the complex architecture of artificial neural networks.

The intricacies of these processes, and of datasets in particular, is explored by Kate Crawford and Trevor Paglen in ‘Excavating AI:The Politics of Images in Machine Learning Training Sets’, to demonstrate how and what computers recognise — and indeed misrecognise — in an image. Computer vision systems make decisions, and as such exercise power to shape the world in their own images, and further reflect existing biases. This problem of bias and the skin/surface is developed by Murad Khan in ‘Notes on a (Dis)continuous Surface’, in exploring ethical questions over the role of automated data-processing instruments, specifically machine learning algorithms, and the role they play in further entrenching existing racial inequalities, racial biases and practices of discrimination. The essay exposes how racial representation functions within machine-learning systems (itself inherently contaminated by the legacies of the colonialism), ‘asking both how race is understood, and what can be achieved by encoding this understanding’. The discriminatory logic of AI is further examined by Leonardo Impett in ‘Irresolvable contradictions in algorithmic thought’, drawing attention to the ongoing contradictions between the commercial interests of Big Tech and the rhetoric of a fairer AI (so-called ‘Responsible AI’) — unable to escape the underlying contradictions at an algorithmic level and in deep learning neural networks.

Following from this, Eva Jäger introduces the Creative AI Lab — a collaboration between the R&D Platform at Serpentine Galleries and King’s College London’s Department of Digital Humanities, and its first project  Database of Creative AI - initiated in 2020 to collect tools and resources for artists, engineers, curators and researchers interested in incorporating machine learning and other forms of AI into their practice. A discussion on Serpentine’s R&D Platform, is further developed Victoria Ivanova and Ben Vickers in their paper ‘Research & Development at the Art Institution’. The text suggests possible directions for extending the discussion to cultural institutions and questions of infrastructure, and to consider what they call ‘future art ecosystems’. An extract from the larger document, the first annual briefing paper called Future Art Ecosystems, is also reproduced here (Chapter 3: ‘Strategies for an Art-Industrial Revolution’)

Returning to some of the discussions around posthumanism, a more subjective register is offered by Magda Tyzlik-Carver in ‘Curating Data: infrastructures of control and affect … and possible beyond’, in which she describes the bodily experience of a curator and writer working with data. She writes: ‘I am sensing how it feels to become posthuman, a body of data and affect.’ As curating becomes increasingly posthuman, it takes place at different levels - it has become an organised form of control executed by algorithms and made possible by big data, while also directly affecting people whose lives have been incorporated into digital infrastructures that maintain the system, a necessary element for the profitable performance of Big Tech.

Finally, we return to the proposition of the title of the journal, ‘The Next Biennial Should be Curated by a Machine: A Research Proposition’, in a text by Joasia Krysa andLeonardo Impett. It introduces a conceptual premise of a larger research proposal that takes the form of various machine learning experiments developed in the context of Liverpool Biennial 2021 to explore machine curation and audience interaction in virtual LB2021.

Stages #9: The Next Biennial Should be Curated by a Machine is edited by Joasia Krysa and Manuela Moscoso. Cover features Manuela Moscoso's curatorial sketch for Liverpool Biennial 2021, one of several sketches drawn during the course of conversations between the editors in connection with the The Next Biennial project.

This volume is produced in collaboration with DATA Browser book series, and will be published as an expanded version in 2021/22 (Open Humanities Press).[4] It has been made possible by the generosity of all contributors, and with the support of Creative AI Lab, Serpentine, London.

[1] Liverpool Biennial 2021: The Stomach and The Port, curated by Manuela Moscoso, 20 March – 6 June, https://www.biennial.com/2020

[2] See the Glossary of terms in this volume, derived from Winnie Soon’s and Geoff Cox’s book Aesthetic Programming: A Handbook of Software Studies, London: Open Humanities Press, 2020.

[3] The Next Biennial Should be Curated by A Machine is a research proposition and an umbrella concept that gathers various experiments exploring the application of machine learning techniques to curating; title and original curatorial concept by Joasia Krysa, technical conceptualisation and development by Leonardo Impett, first experiment B³(TNSCAM) developed as a collaboration with artists Ubermorgen, co-commissioned with the Whitney Museum of American Art for its online platform artport, curated by Christiane Paul. Further research funded as part of UKRI/AHRC Strategic Priorities Fund: Towards National Collection at: ai.biennial.com

[4] See: http://www.openhumanitiespress.org/books/series/da...