Learning to trust the machine: The psychological relationship forming in human-AI language learning.
Date of Presentation
7-29-2026 2:00 PM
College
College of Science & Mathematics
Faculty Sponsor(s)
Dr. Maxi Heitmayer
Poster Abstract
One of the main issues that second language learners face is not having someone to verbally practice with. As a result, people are increasingly learning languages by talking with conversational AI, yet we know little about the psychological relationship that forms through this repeated interaction. This study examines whether intermediate Spanish learners develop an attachment-like bond with a conversational AI voice tutor over three months of daily practice, and whether that bond shapes their willingness to speak. Drawing on attachment theory, the study adapts the dimensions of closeness, dependency, and conflict from educational relationship research to the human-AI context, framing the AI as a potential "secure base" that, combined with Vygotsky's zone of proximal development, may lower the fear of negative evaluation known to inhibit language production.
During this longitudinal, qualitative design, fifteen adult learners will practice speaking Spanish with an AI voice assistant in their everyday environments, recording weekly first-person video of their sessions and completing monthly replay interviews in which they revisit and reflect on their own footage. Proficiency is tracked through an AI-based conversational assessment and anchored by an external, human-scored standardized measure administered at the study's start and end by the World Languages Department. Data will be analyzed using reflexive thematic analysis.
This design addresses a gap in existing research, which has largely examined conversational AI through short-term, outcome-focused, or screen-based studies rather than the lived, relational experience of learners over time. By centering the psychological relationship rather than performance alone, this work aims to inform how supportive, emotionally aware educational technologies are understood and designed.
Student Keywords
psychological relationship, human-AI language learning
Disciplines
Psychology
Document Type
Poster
Learning to trust the machine: The psychological relationship forming in human-AI language learning.
One of the main issues that second language learners face is not having someone to verbally practice with. As a result, people are increasingly learning languages by talking with conversational AI, yet we know little about the psychological relationship that forms through this repeated interaction. This study examines whether intermediate Spanish learners develop an attachment-like bond with a conversational AI voice tutor over three months of daily practice, and whether that bond shapes their willingness to speak. Drawing on attachment theory, the study adapts the dimensions of closeness, dependency, and conflict from educational relationship research to the human-AI context, framing the AI as a potential "secure base" that, combined with Vygotsky's zone of proximal development, may lower the fear of negative evaluation known to inhibit language production.
During this longitudinal, qualitative design, fifteen adult learners will practice speaking Spanish with an AI voice assistant in their everyday environments, recording weekly first-person video of their sessions and completing monthly replay interviews in which they revisit and reflect on their own footage. Proficiency is tracked through an AI-based conversational assessment and anchored by an external, human-scored standardized measure administered at the study's start and end by the World Languages Department. Data will be analyzed using reflexive thematic analysis.
This design addresses a gap in existing research, which has largely examined conversational AI through short-term, outcome-focused, or screen-based studies rather than the lived, relational experience of learners over time. By centering the psychological relationship rather than performance alone, this work aims to inform how supportive, emotionally aware educational technologies are understood and designed.