Provenance
Invisible watermarks on AI images: what they can prove
Quick answer
An invisible watermark is a pattern a generator hides in an image's pixels so that a matching detector can later recognise it, even after some resizing or compression. It is different from Content Credentials, which are signed metadata attached to the file. A detected watermark is good evidence the image came from that generator; no watermark tells you nothing, because most generators and edits leave none.
On this page
Two different ideas: pixels and metadata
Provenance for images comes in two forms. Content Credentials, described in Content Credentials on shared images, are signed records attached to the file; they are readable by anyone with a C2PA tool, but they vanish when the file is re-encoded. Invisible watermarks are hidden in the pixels themselves; they survive many ordinary edits, but only the company that made them can usually check for them.
| Content Credentials | Invisible watermark | |
|---|---|---|
| Where it lives | Signed metadata in the file | A pattern in the pixels |
| Who can read it | Anyone with a C2PA reader | Usually only the watermark's maker |
| Survives re-encoding | Rarely | Often, within limits |
| Records edit history | Yes | No, only the origin |
How pixel watermarks work
Systems such as Google DeepMind's SynthID adjust generated images in ways that are imperceptible to people but detectable by a matching model, and are designed to survive common changes such as cropping, filters and compression. The detector then reports whether the pattern is present, sometimes with a confidence level. The EU AI Act requires providers of generative systems to mark synthetic output in a machine-readable way, which is pushing more generators towards watermarks, metadata or both.
What a watermark result proves
- Watermark found: strong evidence the image, or the part carrying the pattern, came from that generator.
- No watermark found: almost nothing. The image may come from a generator that does not watermark, may have been heavily edited, or may be a photograph.
- Either way: nothing about who made it, why, or whether what it depicts happened.
That asymmetry matters for anyone sharing images. A positive result is useful; a negative one should never be presented as proof that an image is real. The guide to spotting AI images in your feed covers the checks to combine it with, and checking Content Credentials covers the metadata side.
For publishers using generated images
If your images come from a generator that watermarks them, that is a quiet safety net: platforms and fact checkers who can read the mark will recognise the image even after your CDN recompresses it. It does not replace a caption. Readers cannot see a watermark, and the labelling habits in the provenance section still apply.
Sources
- Google DeepMind: SynthID watermarking
- Regulation (EU) 2024/1689, the AI Act (EUR-Lex)
- C2PA, the Coalition for Content Provenance and Authenticity
Questions
What is an invisible watermark on an AI image?
A pattern hidden in the pixels by the generator, imperceptible to people, that a matching detector can recognise later.
Can I check an image for an AI watermark myself?
Only where the watermark's maker offers a public checker. Most pixel watermarks can be read only by their maker's tools.
Can watermarks be removed?
They are designed to survive common edits, but heavy editing, regeneration or deliberate attacks can weaken or remove them.
Is a watermark the same as Content Credentials?
No. Content Credentials are signed metadata anyone can read; a watermark is hidden in the pixels and usually proprietary.
Hannah Voss, Editor. Checks every guide against a working WordPress install and the networks' current documentation. Last reviewed September 2026.
Related reading
- Checking Content Credentials before you share an imageHow to check an image's Content Credentials before you share or publish it: what a C2PA manifest shows, what a missing one means, and what to record.3 min read
- Spotting AI images in your feed: checks that hold upWhich checks for AI-generated images in a social feed actually hold up: platform labels, provenance data, earlier copies and the source, and which tells fail.3 min read
- Content Credentials on shared images: what survives a shareWhat C2PA Content Credentials are, why WordPress image sizes and social uploads often strip them, and how to keep provenance on the images you share.4 min read
- How platforms label AI content, and what triggers the labelHow major platforms decide to label AI-generated or altered media: creator disclosure, provenance metadata, detection, and how publishers avoid mistakes.4 min read