Short answer: AI detectors estimate whether text looks machine-written, and their own makers and independent researchers have shown they can be wrong, including about human writers. Working drafts that you timestamp as you go answer a different, more checkable question: did this exact file exist by this date? That makes a dated trail of drafts a steadier record of your process than any detector score, though it shows when files existed, not who wrote them.
Disclosure: I’m Inna Hall, co-founder of TimeBinder, the file timestamping tool this guide uses. This page is general information, not legal advice.
Table of Contents
What detectors measure
An AI detector reads a piece of text and estimates how likely it is to have been machine-generated, often by looking at how predictable the wording is. It produces a score, not a record. It can’t look at how the text was made; it can only look at the finished words. Some AI companies now also watermark their models’ text, which is a different kind of signal; see what AI text watermarks mean for writers.
What the public record says
It is worth being careful here, because detectors are improving and vary a lot. But a few public sources are clear:
- OpenAI released its own AI text classifier in January 2023 and withdrew it on 20 July 2023, saying it was “no longer available due to its low rate of accuracy” (OpenAI).
- A 2023 study by Liang and colleagues, published in Patterns, tested seven detectors and found they misclassified more than half of a set of essays by non-native English writers as AI-generated, while classifying essays by US eighth-graders accurately (Liang et al., 2023).
- Vanderbilt University disabled Turnitin’s AI detector in August 2023, citing concerns about false positives and how the tool works (Vanderbilt University).
None of this means every flag is wrong. It means a score alone is a shaky basis for deciding what one person did.
What drafts and timestamps show instead
Working drafts answer the question a detector can’t: how did this work come to be? An outline from March, a messy first draft from April and a revision from May tell a story about process that a finished file can’t.
The weak point of drafts is their dates. File dates are easy to change, and cloud version history sits in one account. A timestamp fixes that: it records the SHA-256 fingerprint of each exact file on the Bitcoin blockchain, so anyone can later check the file existed by that time and still matches its fingerprint. See how to prove a file existed before a date and using AI as a writing tool and showing your own work.
To keep track of the drafts in between, Draft Streak logs each day’s word count and file fingerprint; its dates come from your device’s clock, so TimeBind the milestone files for a recorded time.
Prove your draft history. Seal each draft, note and manuscript as you go, so you can show when every version existed. See how it works →
Files are hashed locally on your computer and are never uploaded.
A habit for each milestone
- At each milestone, save a copy of the draft with the stage in the file name.
- Seal it with any notes from that stage.
- Keep the certificate with the files and work on from a fresh copy.
Writers can read more on manuscript timestamping and proving a manuscript existed before an AI accusation.
What a timestamp does and doesn’t do
TimeBinder anchors each file’s SHA-256 fingerprint to Bitcoin and gives you a Proof of Time Certificate. That is verifiable evidence of when a file existed: it shows the file existed no later than the anchor date, and any later change to the file is detectable. ASAP TimeBinds can be checked independently; Pooled TimeBinds are verified through TimeBinder. Your file is hashed in your browser and never uploaded.
It doesn’t show who made the work or who owns it, it doesn’t stop copying, and it doesn’t cover a version you didn’t timestamp. A timestamp doesn’t measure whether AI was used; it shows when each version you sealed existed.
Frequently Asked Questions
Are AI detectors always wrong?
No. They can be right, but they make mistakes in both directions, and a score is an estimate rather than a record. That is why they are a weak basis for a decision about one person’s work.
Can I use a timestamp to dispute a detector score?
You can use it as evidence of your process. A series of timestamped drafts shows how your work developed and when. It does not show who wrote the text, and it does not measure whether AI was used.
Do I need to timestamp every small edit?
No. Milestones are enough: outline, first draft, major revisions and the version you submit. One Pooled TimeBind covers up to 10 files.
Are my drafts uploaded?
No. Each file is hashed in your browser and only the fingerprint is recorded. Your drafts stay on your computer.
Plan details were checked against timebinder.io/pricing on 29 October 2026.
