Research Seminar - Structured Like a Language Model: Analysing AI as an Automated Subject

Structured Like a Language Model: Analysing AI as an Automated Subject

Presented by: Dr Luke Munn

Date: Monday 24 April
Time: 12-1pm
Location: Online via Zoom and in-person at the SCA Writer's Studio (Level 6, Michie Building)

Abstract

The rapid rise of AI language models has drawn various responses, from claims of sentience to dismissals of them as machinic parrots. While acknowledging their limits as technical systems, interacting with them feels like interacting with a subject. For this reason, we turn to a field deeply invested in subjectivity and language: psychoanalysis. We carry out a case study of InstructGPT, detailing its construction and conducting exploratory interviews with chatbots. These investigations draw out how the model is like and unlike human subjects, including short term memory loss, self/object confusion, and competing logics of being helpful, truthful, and not harmful. We conclude that critical media methods and psychoanalytic theory together offer a productive frame for grasping the powerful new capacities of AI-driven language systems. 

Presenter

Luke Munn is a Research Fellow in Digital Cultures & Societies at the University of Queensland. His wide-ranging work investigates the sociocultural impacts of digital cultures, from data infrastructures in Asia to platform labor and far-right radicalisation, and has been featured in highly regarded journals such as Cultural Politics, Big Data & Society, and New Media & Society as well as popular forums like the Guardian, the Los Angeles Times, and the Washington Post. He has written six books: Unmaking the Algorithm (2018), Logic of Feeling (2020),  Automation is a Myth (2022), Countering the Cloud (2022), Red Pilled (2023) and Technical Territories (2023 forthcoming). His work combines diverse digital methods with critical analysis that draws on media, race, and cultural studies.