Affording Creativity with Conversational Agents: How perceived age difference in AI-generated voices influences creative expressions
Authors
Alaoja, Maria Sylvia ; Kristinsdóttir, Guðfinna
Term
4. term
Education
Publication year
2025
Pages
52
Abstract
This Master’s thesis in TechnoAntropology is based on a collaboration between Aalborg University’s Department of Materials and Production, Robotics and Automation team, and LEGO®. It explores how child-like versus adult-like AI generated voices affect verbal creativity in a play-based setting, where we custom-built voiceinteractive game and adapted TTCT metrics. Drawing on both qualitative and quantitative methodologies, we analyse the socio-technical conditions that shape creativity assessment and engagement. We critically examine the assumptions embedded in standardised creativity metrics through the lens of the Sociology of Scientific Knowledge (SSK), and propose a context-sensitive approach to measuring creativity as a socially embedded and co-constructed phenomenon. Preliminary data from adult participants establish a comparative baseline for a forthcoming study with children aged 9–12. Through this interdisciplinary approach, the project contributes to more responsible and reflective AI design for child users.
This Master’s thesis in TechnoAntropology is based on a collaboration between Aalborg University’s Department of Materials and Production, Robotics and Automation team, and LEGO®. It explores how child-like versus adult-like AI generated voices affect verbal creativity in a play-based setting, where we custom-built voiceinteractive game and adapted TTCT metrics. Drawing on both qualitative and quantitative methodologies, we analyse the socio-technical conditions that shape creativity assessment and engagement. We critically examine the assumptions embedded in standardised creativity metrics through the lens of the Sociology of Scientific Knowledge (SSK), and propose a context-sensitive approach to measuring creativity as a socially embedded and co-constructed phenomenon. Preliminary data from adult participants establish a comparative baseline for a forthcoming study with children aged 9–12. Through this interdisciplinary approach, the project contributes to more responsible and reflective AI design for child users.
Documents