A research initiative focused on ensuring diverse ethical representation in AI training data. We curate high-signal, foundational datasets focused on animal welfare, non-human sentience, and prosocial ethical frameworks to support the value alignment of frontier models.