Uncovering AI

Uncovering AI

UKAI Projects · Local Disturbances - Shorts #17 - Uncovering AI

Local Disturbances exists as a means of sharing ideas about how artificial intelligence might be understood, used, and critiqued through the lens of art and culture. We have our own biases in this, and others will certainly have their own ideas about what role we ought to play in shaping the development of algorithmic systems.

The position at UKAI Projects can be summarized as follows:

Most AI is being developed with the values of growth and efficiency (scale) in mind. Things are designed to go faster and bigger. Even in more ‘socially minded’ solutions, the intent is overwhelmingly to optimize otherwise scarce resources. Other values — such as love, community, compassion, or autonomy — are being marginalized or excluded outright through both the technical design of systems and the institutions that determine the direction of development. We are seeking to increase access to algorithmic tools and techniques so that other values and moral positions can benefit from these advances. We also seek to interrogate the hegemony of scale as a moral position in the world around us.

AI ethics too often relies on abstracted arguments that quickly become polarized. We confuse our maps with the terrain. Making sense of AI will require direct engagement with experiences of AI so that we can individually and collectively determine what things “mean”. Ethics can then extend from a grounded experience of sense-making rather than from debate of abstracted (and often Western) principles.

By turning over decisions to institutions and automated systems we forego our own potential for grace and care for others. We lose our position as subjects and quickly are re-framed as objects in need of optimization. We are less concerned with machines that think like humans and more worried about the demands on humans to behave more like machines.

From these positions, we can then go about being curious about how other values can be centered in cognitive technologies, how we can provide embodied experiences of AI to allow others to develop their own point of view from their own position in the world, and how we might develop AI systems that embrace creative interaction and sociality.

The Ferment AI residency

The Ferment AI residency launched in September 2021 with nine participating artists and was an early effort to explore these positions and to ask questions about how to move forward.

In Spring of 2022, the artists participated in an in-person two-week residency at the Queen’s University Ingenuity Labs, an interdisciplinary initiative at Queen's University focused on combining artificial intelligence, robotics, and human machine interaction to create future intelligent systems and robotic machines that enhance human productivity, safety, performance, and quality of life. We had time, space, support, and an existing research community to advance prototypes of AI artworks that were developed through the first five months of the residency. I’m excited to share these projects and how they relate to the three themes described above in the coming months..

What we’re learning

The Ferment AI cohort contains diverse disciplines and different objectives. Selection, however, prioritized artists concerned with the materiality, sociality, and relationality of artificial intelligence and these themes have been the areas where exchange has proven most fruitful. Our collective exploration has resulted in some wonderful, half-baked directions that will require additional exploration (and building).

One emerging theme, at least in my own practice, is around the potential for large natural language processing (NLP) systems to be taught to reify particular ontologies, phenomenologies, and models of space-time as agents in different storytelling approaches.

The most interesting cities, at street level, demonstrate a range of architectural styles and approaches mixed up together. Architecture is a means of making solid a set of assumptions in a given place and time about what is involved in a good life. Le Corbusier made manifest ideas such as “a house is a machine to live in”. The Metabolists in Japan created reminders of an idea that the individual, the house and the city are all parts of a single organism. We see these ideas represented in the buildings around us and their interplay provides a sense of changes over time.

Language is similar. Novels and other forms of literature serve to “freeze” particular ways of knowing the world for exploration and re-animation by others. To Mikhail Bakhtin, each narrative has an “intrinsic connectedness of temporal and spatial relationships” in constructing a particular world. Dostoevsky and Rabelais, as examples, captured moments of collision among theories of time and space in Russia and France, respectively, and part of their ongoing appeal is their ability to bring the reader into those moments of profound upheaval. For Dostoevsky, it was the encroachment of industrialism and modernity (linear time, efficiency, extraction) into older feudal arrangements (circular time, fealty, rootedness, and thrownness).

Might it be possible to train large sets of language data to represent or embody a particular ontology or intersection of ontologies as an agent in a narrative (interactive or fixed). The West is dominated by language that describes ideas of progress (the better future), of competition (metaphors of war and sport), and alienation (abstract nouns, passive constructions). How might current events be described by someone embedded in a different cosmology and with different assumptions about time, space, and the nature of being? How might possible futures describe and encode the events of our volatile present?

The potential here is considerable. It is difficult to critically reflect on one’s own era while embedded in it. Reading Rabelais offers more than a brilliant story, but also insights into how others lived in the past and how they understood the universe around them. Constructed algorithmic agents that can, at least in terms of language, call forth these routines of sense-making can generate novel interactions and complicate our existing and often ossified patterns of thinking and behaving.

Share your thoughts, your curiosities, your ambitions in the comment section and let us know what you’re learning as we collectively figure out the role of AI in supporting better and more beautiful lives.

Back to blog

Leave a comment

Please note, comments need to be approved before they are published.