Guides
What AI watermarking actually is, who does it, and how to check your own text. Written to be accurate rather than alarming.
Law and policy
AI watermarking law: what disclosure rules actually require
The regulations pushing providers to mark AI output, what machine-readable disclosure means in practice, and the gap between marking content and disclosing it.
4 min readWhat C2PA is, and why it is not a text watermark
Content Credentials attach signed provenance metadata to image files. How to check for it, why it is trivially removable, and why text needs a completely different mechanism.
3 min readEU AI Act Article 50: why every provider is watermarking text now
The regulation that became enforceable on 2 August 2026, what it requires, and why Anthropic applied it worldwide instead of only in Europe.
2 min readAnthropic's watermark
How Claude's watermark works
Anthropic has not published its mechanism, but the class of system it belongs to is well documented. What is confirmed, what is inferable, and what is guesswork.
4 min readDoes Anthropic watermark Claude's text? Yes — since August 2026
Anthropic now embeds an invisible watermark in Claude's text output. What it is, what it proves, and what Anthropic itself says degrades it.
4 min readCan you remove Anthropic's text watermark?
An honest answer: what degrades a statistical watermark, what cannot touch it, and why nobody can currently verify the result.
3 min readDetection
Is Claude detectable?
Three different questions get asked as one. Whether Claude's watermark can be read, whether AI detectors flag Claude text, and whether anyone can prove you used it.
3 min readHow AI detectors actually work
Perplexity, burstiness, and why every detector reports a probability rather than an answer.
2 min readWhy AI detectors flag human writing
The false positive problem is structural, not a tuning issue. Who it hits, why, and what to do if it happens to you.
2 min readRemoving the watermark
How to remove ChatGPT's watermark from text
OpenAI researched a text watermark and did not ship it. What that means for GPT output, what is genuinely in it, and what a 'ChatGPT watermark remover' is actually removing.
3 min readHow to remove Gemini's watermark from text
Gemini text carries SynthID, the longest-running text watermark in production. What is in Gemini output, what can be stripped exactly, and what only a rewrite touches.
3 min readCan you remove SynthID from text?
SynthID is the best-documented text watermark there is, which makes it the only one where removal can be discussed precisely. What degrades it, what cannot touch it, and what you can verify.
4 min readDoes Claude's API output carry the watermark?
For developers: whether text from the Anthropic API is marked, which models, and what it means for products built on Claude.
2 min readHow to check if text has Claude's watermark
What you can detect yourself, what no public tool can detect, and why every 'Claude watermark checker' claiming a percentage is guessing.
2 min readClaude's watermark and Turnitin: what schools can actually see
Whether plagiarism checkers can read Claude's watermark, what they do instead, and why a flag is weaker evidence than it looks.
2 min readHow to remove Claude's watermark from text
A step-by-step guide to removing Claude's watermark: what can be stripped exactly, what needs rewriting, and what nobody can verify yet.
4 min readIs removing Claude's watermark allowed?
What the EU AI Act actually requires of users versus providers, what institutional rules cover, and where the real risk sits.
3 min readRemove Claude's watermark for free
Which parts of watermark removal are genuinely free, which cost money and why, and how to do the free part yourself.
2 min readRemoving Claude's watermark in Word, Google Docs and email
Where hidden characters survive, where they get stripped for you, and how to clean text without breaking your formatting.
3 min readWhich providers watermark
Claude vs ChatGPT vs Gemini: which AI watermarks text
Where each provider stands on text watermarking as of August 2026, what is verifiable, and what it means for output you already have.
2 min readDoes ChatGPT watermark its text?
What OpenAI has researched, what it has actually deployed, and what remains detectable in GPT output.
2 min readWhat SynthID-Text is and how statistical text watermarking works
Google shipped the first text watermark and documented it. Understanding SynthID explains how Anthropic's new Claude mark most likely behaves too.
2 min readHidden characters
Zero-width characters: the invisible marks that really are in your text
ZWSP, ZWNJ, ZWJ, word joiners and the BOM. What they are, why they survive copy-paste, and how to find every one.
2 min readUnicode tag characters: how a whole message hides inside plain text
The U+E0000 block renders as nothing anywhere. It can carry an entire hidden payload attached to ordinary words.
2 min readVariation selectors: the second invisible carrier
U+FE00-FE0F and the supplement block ride invisibly on ordinary letters. A quieter hiding place than zero-width characters.
1 min readHow to audit a document for hidden characters
A practical checklist for checking any text you did not write yourself, and for verifying that cleanup actually worked.
2 min readStylistic markers
Why AI text is full of em dashes
The em dash is the most reliable typographic tell in model output. Why it happens, and why replacing it is safe.
2 min readThe words that give away AI writing
The vocabulary that appears far more in model prose than human writing, why it clusters, and what to do about it.
2 min readGuides
Want to know what is in your own text? The scan is free and runs in your browser.
Scan my text