To propose that GenAI can compensate for "archival silences" perverts Trouillot's work. It's is a contradiction to say that a product whose logic exacerbates power imbalances--creating a most-likely guess based on the existing record--can restore absences that are the product of those imbalances.
