Products & Tools · 2025
FitnessTracker - Ideal Self
A future self, made tangible.
An experimental platform exploring how AI-generated self-modeling can support motivation and long-term behavior change in fitness.
Overview
FitnessTracker is an experimental web platform that explores how AI-generated self-modeling can support behavior change in fitness contexts. The system allows users to select an ideal body shape, generates a personalized future-self video, and integrates it into a lightweight tracking and reflection interface.
It served as an early prototype for investigating how personalized AI representations influence motivation, perception, and adherence over time. The work informed later research directions on engagement decay and personalization.
Design notes
Can a visualized future self change what we do today?
FitnessTracker investigates the gap between abstract fitness goals and the daily actions needed to reach them. Instead of relying only on numbers, streaks, or reminders, the experience gives each person a vivid representation of a possible future body.
The platform is a continuity of a paper published in CHI'26 (Checkout Here). It treats AI-generated ideal self as a behavioral interface: personal enough to feel relevant, but lightweight enough to live inside an everyday tracking routine.
From aspiration to reflection
People choose an ideal body direction, generate a personalized future-self video, and connect that image to a simple loop of tracking and reflection.
The prototype was designed as an early research instrument. It helped surface questions about motivation, perception, novelty effects, and whether personalized nudges remain meaningful over time.