fairfolio
Saved Projects

A Novel Computational Framework Using AI Agents as Synthetic Participants to Study Avatar Self-Representation Behavior

CSEF · 2026 Behavioral & Social Sciences (Senior Division)

Nisa Thomas

Overview

Digital avatar-based identity is rapidly becoming central to adolescent life, with 85% of Gen Z users having created a digital avatar. As individuals increasingly represent themselves through digital bodies, avatar self-representation behavior may encode meaningful psychological signals, offering a potential pathway for non-invasive behavioral research on body image and related mental health concerns. However, studying this domain with human participants is ethically constrained, costly, and time-intensive, leaving this emerging behavioral frontier largely inaccessible. This project introduces a novel computational framework that uses AI agents as synthetic participants, each assigned a unique psychological profile derived from validated human survey data, to run controlled behavioral experiments on digital avatar self-representation ethically, efficiently, and at scale. The framework integrates four components: controlled 3D avatar stimulus generation, computer vision feature extraction, psychological profiling using Big Five personality traits and Eating Disorder Examination Questionnaire behavioral dimensions, and structured agent-based behavioral simulation. Using this framework, 5,600 controlled trials were conducted to analyze avatar switching behavior across systematically manipulated body features. Results demonstrated that switching decisions were primarily driven by body perception cues, while eating disorder risk traits modulated sensitivity to these features. A perceptual control condition using inverted avatars significantly reduced switching behavior, confirming that the observed patterns reflect meaningful behavioral responses rather than random outputs. These findings align with established body image literature, supporting the validity of the framework. These findings demonstrate the feasibility of AI-agent simulations as a new scientific tool for behavioral research, enabling the study of sensitive and previously inaccessible domains. More broadly, this framework introduces a scalable and ethical pathway for investigating digital self-representation and its relationship to psychological well-being, establishing a foundation for future non-invasive tools for studying body image perception and eating disorder risk in digital environments.

Competition history

  • CSEF 2026 Behavioral & Social Sciences (Senior Division) · Entry S-03-22

Related projects

Closest projects by meaning, across every fair and year in the corpus.

Browse more like this

Source: California Science & Engineering Fair public projects

Save projects to your library

Sign in with Google to keep track of projects you find interesting, organized into folders. An account also raises your daily allowance for “Has this been done?”, and lets you create a key for the MCP server with a much higher limit than anonymous use. Browsing stays public.

Continue with Google