Reproduction is costly. Despite this, evidence suggests that parents sometimes feed unrelated offspring. Several hypotheses could explain this puzzling phenomenon. Adults could feed unrelated offspring that are 1) of their close social associates to facilitate these juveniles’ integration into their social network (the social inheritance hypothesis), 2) potential extrapair offspring, 3) at a similar developmental stage as their own, 4) coercing feeding by begging, or 5) less-developed (to enhance their survival, which could benefit the adult or its offspring; the group augmentation hypothesis). Colonial breeders are ideal for investigating the relative importance of these hypotheses because offspring are often kept in crèches where adults can exhibit allofeeding. Using automated monitoring of replicated captive zebra finch (Taeniopygia guttata) colonies, we found that while parents selectively fed their own offspring, they also consistently fed unrelated offspring (32.48% of feeding events). Social relationships among adults prior to breeding did not predict allofeeding, nor was allofeeding directed toward potential genetic offspring. Instead, adults with more-developed offspring preferentially fed less-developed non-offspring over non-offspring at a similar developmental stage as their own offspring, and this tendency was not explained by differences in begging behavior. Our study suggests that allofeeding is consistent with group augmentation, potentially benefiting adults through colony maintenance or increased offspring survival.