Our paper was accepted to the Pluralistic Alignment and Trustworthy AI for Good workshops at ICML 2026.