Bio
Jana Thompson is currently a grad uate student in design at Maryland Institute College of Art, holds an MA in Germanic Studies, BS in mathematics, and BA in anthropology from the Univers ity of Texas at Austin, and has held fellowships in cultural studies in Berlin and data science in San Francisco. She was until August 2020 a member of Los Angeles- based Feminist.AI, where she created workshops on s ocio- technical approaches to AI and developed AI approaches to XR and mental health. Recent events have included Embodied Ethnography: Posthuman and Cultural Approaches at Vienna Anthropology Days 2020, a workshop on Cultural AI design at IxDA 2020 in Milan, and a poster presentation on Feminist.AI project Contextual Normalcy at the Frontiers of AI in Health Care Symposium, held at Stanford University in Fall 2019.
Communication
Abstract
Large scale encyclopedic approaches to artificial intelligence have had the goal of creating a truly intelligent agent. These projects have their intellectual roots in the computational works of Ramon Llulland Gottfried Leibniz in the creation of mathematical systems that proposed to encompass and generate all possible knowledge. The encyclopedic all-knowing god of AI that today manifests itself in the quest for artificial general intelligence –the omniscient AI that is the end-goalof projects such as American-based organizations Cycorp and OpenAI. In recent work, Bender and Koller (2020) challenge these projects in grand claims of understanding such as OpenAI’s GPT-2 in the realm of natural language understanding. Furthermore, the work of Searle (1980), Adam (1998) and others indicate the postulationthat these projects could in the end produce nothing more than a constructionbest represented by Microsoft’s Taychatbot –an amalgamof expressed human opinions in misogyny, racism, transphobia, and other bigotrieswithout understanding of its own.
In stark contrast, there has been a different approach on AI as memory and storyteller. This approach is exemplified in the work of Eugenia Kuyda’s Replika and Stephanie Dinkins N’TOO, which serve as repositories of those they have loved and wish to memorialize. Their workscreatean extension of the human into the non-biological and computational space, creating work that is both approachable and meaningful in the human-computational dialog (see Hayles 2005, 2006). By situating their work within the highly personal, they root their approaches to AI within a subjective meaning, allowing for extension and existence that is both dynamic, evolutionary, and significantforthe humans who interact with their work. Indeed, people have reported ongoing conversations with Replika that resembled that of a person with their therapist and have credited Replika with saving their lives. N’TOO, was created “to ensure that people of color, and others who inherently understand the need for inclusion, equity, ethics, and multimodal testing, participate in the design, production, and testing of ‘smart’ products’. The creation of AI as a labor of love, of the disembodied minds of others, reimagines the storyteller in the posthuman space, as extensions and unique creations.
In contrasting these two approaches, it is possible to examine AI withboth epistemological and semiotic lenses to shed greater light on systemic approaches to technological advances that increasingly pervade our lived experiences, our shared stories, and our possible future.