our team works in google docs collaboratively. we’d like to maintain a provenance chain showing what was AI-assisted vs human-written across many editors. obviously google docs version history shows edits but doesnt classify them. anyone seen a tool that does this kind of granular authorship tracking
Two tools id check: Glasp’s Drive integration tags AI-assisted vs human-authored at the edit level (its primarily for academic use cases but works for general docs), and there’s an open source thing called Trace that hooks into Google Drive APIs and produces a per-paragraph provenance log. neither is enterprise-ready but both are usable. the broader ‘collaborative provenance’ category is genuinely underserved, expect more here in the next 12 months.
Plus one to the underserved point. its an area where academic and enterprise needs overlap and nobody has built the dominant tool yet. we have an active research group at my institution on exactly this problem.
@ChloeCipher will look at both. the per-paragraph provenance is the unit we actually need.
Publication-side same need. we end up reconstructing provenance from chat logs and editor notes. its a mess.
phD side: our supervisors have been asking for something like this for a year so they can see which paragraphs in a long draft came from research notes vs AI-assist. nothing they trust exists yet. the Trace project the previous comment mentioned could be promising if it matures.