Mapping value(s) in AI

Mapping value(s) in AI

 

This interdisciplinary project puts mechanisms of value production at the center of considerations about the design and effectivity of AI solutions. Values represent attributions or perceptions of relative worth, merit or importance. Rather than asking which values AI should satisfy, the project addresses how values ‘map’ between computational and context-sensitive social processes. The project investigates how values are ‘assembled’ in ecologies of human and computational agents out of translations from – e.g., economic and informational models, social theories, ethical principles – to define and justify design goals and decisions. An investigation framework will be prototyped for a recommender system, targeting terms/tokens of value discoverable in design processes as well as background conditions structuring their use. The project asks: What conditions make specific terms/tokens of value accessible, intelligible, and attractive for design? How/do terms and tokens of value change across networked deployments? What work do assembled values do once deployed?

 

 

Bernhard Rieder is Associate Professor of New Media and Digital Culture at the UvA. His research focuses on the history, theory, and politics of software.

 

 

 

Geoff Gordon is a Senior Researcher in international law researching value production and technological governance on the basis of science and technology studies, network theory, and critical legal method.

 

 

 

Giovanni Sileno is a Postdoc at the System and Networking Lab of the Informatics Institute at the UvA investigating computational regulatory systems and artificial cognition.