Dislike-Shaping Dating Explorer (GPTs)
Find your ideal partner by eliminating what you dislike
Overview
A GPT-powered dating preference exploration tool developed for Prof. Hwajung Hong's "AI & Society" course at KAIST Department of Industrial Design. The core insight: when people select their ideal partner by choosing preferred traits, they end up with an unrealistically perfect person. Instead, this system asks users to pick the person they dislike from two profiles, then derives the complement of disliked traits as the ideal — producing more realistic, attainable partner profiles.
Approach
Built as a Custom GPT using iterative profile generation and elimination. Users repeatedly choose the less desirable profile from pairs, and the system analyzes rejection patterns to surface unconscious preferences through the complement set of dislikes.