How to Prove You Wrote Something
Retroactive proof: what you already have
If you have already submitted and been flagged, your evidence is whatever you recorded during the process. Here is what to look for, ranked by persuasive strength:
Google Docs version history
Your entire edit history with timestamps. Every save point shows how the document evolved. Access: File → Version history → See version history. Screenshot the full timeline and the earliest versions.
Draft files and autosaves
Multiple saved versions showing the document's progression. Check OneDrive (Word), Google Drive, iCloud Drive, or your local Downloads folder for earlier exports you may have saved.
Research trail
Browser history, bookmarks, notes, PDFs you downloaded, tabs you saved. The intellectual work that preceded the writing. AI generation leaves no research trail.
File timestamps
Date Created and Date Modified visible in file properties. Screenshots of file metadata show active editing over time. Right-click → Properties (Windows) or Get Info (Mac).
Emails and messages
Emails to instructors asking questions, messages to classmates about the topic, tutor feedback — all show you were engaged with the material during the writing period.
Why retroactive evidence has limits
Retroactive evidence — file timestamps, version history — is strong but not tamper-proof by definition. A sophisticated institution might argue that these records do not conclusively establish authorship, only access. They can be questioned.
More practically: not everyone uses Google Docs. Not everyone saves drafts. Not everyone has search history enabled. If you have none of the above, your appeal rests on the statistical limitations of the detection tool — a valid but weaker position.
The proactive approach — recording the session as it happens — produces evidence that is structurally different. It was created during the writing, before any question of authorship existed, by a system that had no reason to fabricate it.
Prospective proof: before you are ever accused
The most robust way to prove human authorship is to create evidence at the time of writing — before any dispute begins. This is what Scripli does.
Scripli observes your writing session in real time — recording behavioural signals like writing rhythm, revision patterns, and session timing — and seals that evidence into a tamper-proof cryptographic record. The result is a Human Authenticity Certificate (HAC): a signed, verifiable proof that a specific human wrote this document, in a specific session, on a specific date.
A HAC is not an AI detector result. It does not analyse the output and estimate a probability. It certifies that the process occurred. The difference is fundamental: a certificate from the time of writing cannot be contradicted by a detector's analysis of the finished text.
The key distinction
AI detector result
Analyses the output after writing. Produces a probability estimate. Can produce false positives. Created after the fact.
Human Authenticity Certificate
Records the process during writing. Produces a cryptographic certificate. Cannot be falsely generated by AI. Created before any question is asked.
Prove it before you ever have to
Scripli records your writing session and issues a Human Authenticity Certificate. Proof exists before any question is ever asked.