AI alignment should coordinate many viewpoints, not optimize one set of values

Original: Socially Grounded Agentic AI: Coordinating Plural Perspectives through Social Theory

Source: arxiv.org ↗