Skip to Main content Skip to Navigation
New interface
Conference papers

Liens entre confiance et acceptabilité dans un dispositif IA

Abstract : Several studies have presented trust as crucial to predict AI acceptability. We aim to involve statistical and social measures of trust, to confirm links between trust and acceptability, and between trust and emotional variation, with an experimental protocol inspired by [1]. Sixty participants are asked to estimate ages on portraits; then, an AI model suggests its own prediction based on facial features, along with its stated confidence. Participants can then keep or change their initial prediction. Three measures are used to quantify their confidence in the model: the proportion of agreements and changes, and fixation duration. Participants are also asked to complete a questionnaire to measure their trust and acceptance, and their facial expressions are recorded during the experiment to assess emotional variation. Therefore, this research allows to better understand links between trust, emotions, and acceptability in AI.
Complete list of metadata

https://hal.archives-ouvertes.fr/hal-03789503
Contributor : Alexandre Agossah Connect in order to contact the contributor
Submitted on : Tuesday, September 27, 2022 - 2:41:00 PM
Last modification on : Friday, September 30, 2022 - 4:15:57 AM

File

IHM_22___Atelier_IHM_et_XAI (3...
Files produced by the author(s)

Identifiers

  • HAL Id : hal-03789503, version 1

Citation

Alexandre Agossah, Lucie Lévêque, Matthieu Perreira da Silva, Patrick Le Callet, Frédérique Krupa, et al.. Liens entre confiance et acceptabilité dans un dispositif IA. 33ème Conférence Internationale Francophone sur l'Interaction Humain-Machine (IHM 22), Apr 2022, Namur, Belgique. ⟨hal-03789503⟩

Share

Metrics

Record views

12

Files downloads

2