Publications / Human-AI interaction · Self-presentation
AI-generated self-interpretations
How people articulate themselves, and anticipate presenting themselves to others, when they read AI-generated interpretations of who they are.
The research question
What happens when an AI system offers people an interpretation of themselves before they present themselves to someone else?
This paper comes from my master’s thesis at National Taiwan University. It follows Beyond Adaptation, moving from generating ideas for self-presentation to how people respond to seeing themselves described by AI.
Research approach
- 01 · BuildGenerative AI systemTurns personal material into tags, narratives, and images of the self.
- 02 · Study20 participantsThink-aloud lab sessions with emerging adults.
- 03 · AnalyzeThematic analysisOf what participants said as they read the AI’s interpretations.
I developed a generative AI system that turns personal material into tags, narratives, and images of the self, and ran a qualitative lab study in which twenty emerging adults thought aloud while reading what the system said about them.
The analysis describes four overlapping ways participants articulated themselves in response: differentiating, connecting, contextualizing, and comparing. It also separates three things that are easy to treat as one: articulating oneself, accepting an AI output, and anticipating presenting it to others.
Status
The manuscript is under review. Detailed findings will be added here once the review is complete.
My role
- System development
- Study design
- Think-aloud sessions
- Thematic analysis
- Writing
Carried out with my advisor, Nanyi Bi, at National Taiwan University.
Citation
Chien-Yin Wu and Nanyi Bi. A study of AI-generated self-interpretations and anticipated self-presentation. Manuscript under review (title withheld).