Course Description
Truth in an organization is a human judgment about which sources are authoritative. AI systems can be built to amplify truth, but generative AI by itself does not. This course looks at how content design shapes what AI systems produce, and how communications leaders can keep authority and accountability visible.
Students will get hands-on experience building a content repository using the building blocks of machine-readable content: Markdown for documents, YAML for metadata and GitHub for version history. Each student picks a domain (product content, brand messaging or internal comms) and writes the source material themselves, so they know it well enough to recognize when an AI gets it wrong.
Weekly experiments run the repository through AI tools to see where truth lives in each system and how structure, metadata, vocabulary and provenance change the answers. Students write evaluation criteria to judge those answers against a defined standard, and practice the human judgment that automated scoring cannot replace. Readings in media theory and the politics of AI place this work in a longer history of who controls information and how it spreads.
University of Washington
Department of Communication