Human vs. AI – Diff-based line-level provenance for text under agentic editing

A tool called "Us vs. Them" uses git version history to identify which lines of text were written by humans versus AI agents, assigning provenance scores to help agents understand which sections should not be edited. The system works by analyzing diffs across file versions without requiring special markup, producing ranges that show human-authored "islands" within machine-generated "seas" of text. This addresses the practical need in AI-assisted coding to prevent agents from overwriting human-written or human-edited content while allowing them to freely modify AI-generated sections.