Group Meeting: Incorporating Dialectal Variability for Socially Equitable Language Identification