Ruthanne Huising: From Loops to Assemblages: The Generation of Relational Expertise

With this new colloquium series, we aim to create a space for academic exchange, discussion, and networking around topics related to the reorganization of knowledge practices in the digital world. We regularly invite experts who address these issues from a variety of interesting perspectives.

The event series is open to all researchers at the institute as well as interested external participants.
 

Participation

The talks will take place at the Weizenbaum Institute (Flexraum). No registration is required for in-person attendance.

Hybrid participation via Zoom is possible. To do so, please send us a brief request by email to: : kolloquium.rvw[at]weizenbaum-institut.de.

Lecture

Ruthanne Huising (ESSEC Business School Paris): From Loops to Assemblages: The Generation of Relational Expertise (EN)

Abstract

Social studies of ML technologies have largely adopted the technological frames of designers and promoters of these technologies: examining technology as a competitor or collaborator with human cognition, adopting terminology such as domain expertise and human in the loop, and assuming that humans have few possibilities beyond accepting, rejecting, or questioning technological recommendations. Drawing on an eighteen-month ethnography, we examine how gastroenterologists integrate ML technologies into their work, an assemblage of actors – human and non-human – through which they reason and develop knowledge about abnormalities in patients’ colons. We show how the contributions of ML technologies to knowledge production processes varies – even within a colonoscopy – in relation to temporal and epistemic conditions. This study aims to reorient how the introduction and use of ML technologies is conceived of and articulated in several ways. First, we demonstrate that how members of professions produce expertise is never entirely a cognitive act or a human act, showing how expertise is generated relationally and within a system of actors. This challenges dyadic notions of competition or collaboration with ML technologies. Second, we show that when situated and evaluated in relation to the contributions of numerous other actors, questions about ML technologies’ explainability, accuracy, and certainty are largely absent. This challenges ideas that the use and value of ML technologies can be altered through improvements in transparency and performance. Finally, by identifying how the contributions of the technology relate to temporal and epistemic conditions, we provide practical knowledge of the technology in use.

Afterwards, Katharina Berr (Weizenbaum Institute) will comment on the talk before we move on to a joint discussion and professional exchange.

All previous dates of this colloquium series can be downloaded here.