A Is For Another: A Dictionary Of AI

Notícies

Original website here

A Is For Another answers the question: what is AI? Or, how do we understand what it means to be human and non-human through artificial intelligence? This dictionary presents how intelligence, humans, machines, data and mind exist across a variety of cosmologies, sciences and art practices as perceptions change over time.

This project 

This project focuses on a particular aspect of AI: that it is shaped in terms of the human. So while AI is about machine learning, cognition and robots, it is also about how we understand what it means to be human, how our notions of the human, and of the non-human, are changing over time, as are our inter-relationships. We want to show that ‘intelligence’, ‘humans’, ‘machines’, ‘data’ and ‘mind’ exist differently across a variety of cosmologies, sciences, and practices. A Is For Anotherputs diverse scholarship, art and cultural work in conversation with each other to further this thinking.

Artificial Intelligence, ‘AI’, is a suite of technologies that includes machine learning, computer vision, reasoning, and natural language processing, among others. It exists in an awkward and unique space as technology, metaphor and socio-technical imaginary

In 1987, a computer was considered ‘intelligent’ because it beat the world’s highest ranked chess player. Now, a computer will be considered intelligent if it can drive a car; or because it generates a response based on your facial expression. Intelligence becomes a moving target defined by tasks like perception, pattern recognition, language processing or prediction. What is served by this idea?

Now, AI’s component technologies are employed in object recognition, language translation and recommender systems, among others, creating novel manifestations of existing personal, social and political relations.

Metaphors and narratives serve as shorthand for navigating new situations and technologies. The history of science shows us that we use metaphors to describe something unfamiliar and new, and this becomes entangled in how we study it, thus shaping the artefact in terms of its metaphor. Phrases like ‘information just wants to be free’ or ‘data is the new oil’ have shaped what we think digital technologies are. Such metaphors determine how a technology is developed, built, and will be regulated (or not regulated). For example, if you imagine artificial intelligence technologies as a child, a non-human species like an octopus or a mythical creature like a Centaur, then how will you regulate it?

A Is For Another is a response to the world-making and culturally homogenous approaches to AI emerging from places like Silicon Valley and Hollywood. For example, in the European/US/UK Science Fiction genre, robots generally convey an anxiety about the displacement or erasure of human beings. The beautiful, calculating fembot Eva in Ex Machina or the threatening synthetic intelligence Skynet in Terminator are such compelling fantasies.

The interaction between pop culture, cosmologies and the robotics industry in Japan presents a different kind of ‘robotic imaginary’ (Rhee). And ‘robot’ is only one word for this fascinating and dramatic non-human/human being. Tokyo Cyberpunk presents a panoply of vivid posthuman pop-cultural figures negotiating speculative socio-political scenarios, “explor[ing] new possibilities of becoming at the rhizomatic intersection of different forms of intelligence, corporeality, and data processing.”

Also, with Jennifer Rhee we ask of robots and all things AI: “How does this figure engage the histories of under- valued, devalued, and exploited labor, particularly as they intersect with race, gender, and class? Who is being dehumanized in this robot figure? Whose humanness is constructed as familiar and sacrosanct?” So what happens when we read a history of intelligence from the perspective of Cognitive Psychology alongside a social-materialist inquiry into the women who were known as computers?

Or consider that mushrooms and robots, both non-human entities, are equally subjects of the Posthumanities; Anna Tsing’s ethnography of Matsutake cultivation has been an inspirational text for many entries here. However, Alexander Weheliye and Sylvia Wynter have written very different histories of ‘non-humans’, and the category of ‘man’, through the lens of Transatlantic slavery, colonialism and colonial plantations. Similarly, futurisms- Afro-futurismGulf-futurism- are syntheses of geopolitics with pop-culture, and psychogeographies of bodies moving forwards and back in time and space, both online and off. What imaginaries of automation, bodies, data and intelligence exist here? 

The purpose of this project is not to advance a new theoretical rationale for how AI could or should be designed. Nor does it describe how technologies like machine learning or computer vision work in the world, how they demonstrate bias, if they have ethics or not, or might be used to address the climate catastrophe. There are many resources about AI on these topics online. 

Instead, we want to create forks and distractions in how ‘AI’ is being imagined and produced in the world.

This Website

A traditional online search assumes that a seeker already knows what they are looking for and can type it into a box. But as Mitchell Whitelaw tells us: "Search is ungenerous: it withholds information, and demands a query...” His generous interfaces approach suggests an alternative: “rich, browsable interfaces that reveal scale and complexity.”

Padmini and Pratyush — the designers of this website — applied Whitelaw’s approach to reflect my own journey of learning about AI’s [other] dimensions and futures: meandering, serendipitous, non-linear. This process has involved reading about the history of AI as documented in official archives; visiting art exhibitions and cultural events and attending lectures; interacting with futurologists imagining autonomous cars through animal intelligence, and so on. It has been both bewildering and illuminating — and not all connections are strong; sometimes the path is more interesting than where you end up. 

This process has primarily involved searching online; but using the internet for research is a tricky exercise. There are paywalls (and ways to get around them, sometimes), a structural politics to who creates knowledge online, and closely-guarded, proprietary algorithms curating what will appear higher up in your search results. Human curators and guides have been essential to my wayfinding by making material available, and suggesting connections to domains of knowledge not all of which appear related to each other at first glance. 

What do these journeys look like when transposed onto a digital experience? 

The grid view is like the familiar experience of consulting a dictionary: solitary, specific and top-down. It comprises curated, unique and whimsical entries offering different perspectives on humanity and AI. Each entry is a body of text with hyperlinks to references outside this website. 

The relational view is a visualisation of all these references together. Every reference’s hyperlink is assigned a handful of tags; for example: ‘non-human’, 'ecology', ‘posthumanism’, or ‘robots’. These tags are visualised as yellow bubbles; the larger the bubble, the more references associated with that tag. Clicking on a tag bubble reveals smaller white bubbles that link out to that reference. The tag bubbles themselves are not clickable. Pratyush has programmed various cues into the visualisation to orient you to your journey through the material. The visualisation works best when viewed on a desktop screen.

Comparing the grid and relational views, we might say that the grid view is like swimming in a pool, and the relational view is like swimming out into the ocean. In the words of Anna Tsing, this is not a “logical machine” but an “open-ended assemblage”; it “gesture[s] to the so-much-more-out-there.”

This dictionary is neither exhaustive nor comprehensive. It is a snapshot of small and situated data aggregated over years of summer schools, workshops and seminars. This dictionary is not closed either; it is open to material that takes this thinking to new places. There is space for new entries, collaboration and further research and pedagogy, so please get in touch: hello at aisforanother dot net

Maya Indira Ganesh. Berlin, April 2020

This Project

This Website

Grid View

Relational View

Credits

Maya Indira Ganesh. Berlin, April 2020