[ATHENA] Séminaire du CIS avec Jessica Feldman, jeudi 13 juin 2024, 11h à 12h30, à Paris et en visioconférence

[Céline Vaslin <celine.vaslin AT cnrs.fr>]

Bonjour,

Le séminaire du Centre Internet et Société (CIS-CNRS) reçoit Jessica Feldman (American University of Paris), jeudi 13 juin 2024, de 11h à 12h30, à Paris 17^e et en visioconférence.

Présentation : https://cis.cnrs.fr/sem-cis-40-jessica-feldman/
Inscription : https://cis.cnrs.fr/extras/inscriptions/sem_cis.php

Collective Trust? Comparative Commons-Centric Design for AI and Algorithmic Trust

    This seminar introduces a new research project investigating the
    concept of “collective trust” vis-à-vis AI and algorithmic
    decision-making. Based on fieldwork and interviews with horizontal
    and decentralized community groups and social movements, Jessica
    Feldman will present some observed problems arising with the use of
    algorithmic digital security and digital democracy tools. She will
    then outline preliminary reviews of the state-of-the-industry for
    trust in digital design, and of literature on existing theoretical
    paradigms of trust which undergird digital models. 
    
    The groups studied are often stymied in their democratic practices
    at the layer of digital security, where power bottlenecks around the
    holder of the passwords or the managers of the social media
    accounts, and struggles ensue. More challenges arise as they seek to
    scale up across distances, work remotely, or automate decisions,
    requiring the use of digital tools for a form of deliberative
    decision-making that generally does not rely on binary or predictive
    logic. This often leads to serious rifts within the groups,
    disrupting their dynamics of trust and processes of governance. 
    
    In computer science, the problem of trust in AI has arisen as a
    question of accuracy in prediction, or transparency about decision
    making processes, rather than something communal, dynamic, or
    relational. Most existing research on public trust in AI largely
    focuses on developing methods to explain AI in order to encourage
    adoption of the technology by individual users, and to encourage
    public acceptance that AI will be used “on them” in the realms of
    policing or governance service provision. Research on “Trust and
    Computing” also tends to conceptualize the trust relationship as a
    dyadic one between an individual human user and technology.  
    
    These perspectives evidence themselves in the fact that very few
    designs exist that imagine the user as a collective rather than an
    individual. This research project will attempt to articulate what
    algorithmic decision making tools would look like if conceived to
    support and respect “collective trust” and solidarity in democratic
    groups.
    
    Jessica Feldman is a researcher and an
    artist. She is an Assistant
    Professor at the American University of Paris in the Department of
    Communication, Media, and Culture. Before that, she was a
    Postdoctoral Fellow at the Digital Civil Society Lab at Stanford
    University, after earning a Ph.D. in Media, Culture, and
    Communication from New York University.
    
    Her research focuses on the intersection of new information and
    communication technologies and social justice, mainly by connecting
    ethnographic fieldwork with social movements and design
    collaborations with engineers. Her current book project articulates
    core design priorities that emerge in alternative computing tools
    developed by grassroots social movements focused on values of
    democracy, solidarity, and inclusive participation. Her artwork
    addresses similar issues by experimenting with emerging technologies
    and practices of listening. Her research has been published in
    English and French in international journals and edited volumes and
    her artwork has been shown internationally in venues such as
    galleries, museums, concert halls, public parks, and city streets.

Bien cordialement,

***
Centre Internet et Société
CNRS UPR 2000 / GDR 2091
https://cis.cnrs.fr