What is SynthIDAn invisible watermark Google adds to content made with its AI tools

TimeBinder adds a layer of protection against AI and human fraud and disputes.

The file stays on your device. Your browser makes a digital fingerprint of the file. TimeBinder records the fingerprint and the time, not the file.

Short answer

SynthID is Google DeepMind’s technology for adding an invisible digital watermark to content made with certain AI tools, covering images, video, audio and text. A matching detector can check whether a file carries that watermark. It only covers content made with participating tools, so no watermark does not mean a file is real.

How SynthID works

According to Google DeepMind, SynthID embeds digital watermarks directly into AI-generated images, audio, text or video. The watermarks are imperceptible to humans but can be found by SynthID’s own technology. For images and video, the watermark is added when the content is created and is designed to survive changes such as cropping, filters, changed frame rates or lossy compression. For audio made with Google’s Lyria model or NotebookLM’s podcast feature, it is designed to survive things like added noise, MP3 compression or a change of speed. For text from the Gemini app and web experience, SynthID adjusts the word-probability scores used during generation (Google DeepMind).

Google says anyone can now upload an image, video or audio file to the SynthID Detector to check whether it was made with AI from Google or partners it names, including OpenAI, NVIDIA and Kakao (Google DeepMind). Google’s developer documentation says text watermarks can hold up to some edits but that detector confidence can drop a lot if the text is thoroughly rewritten or translated, and that SynthID Text is not designed to stop determined bad actors (Google AI for Developers).

In the EU, providers of generative AI systems must mark synthetic audio, image, video or text output in a machine-readable way (AI Act, Article 50(2)). Watermarks are one way to do that. See does AI leave a watermark and what are Content Credentials (C2PA).

What a watermark check can and cannot tell you

A SynthID result speaks only about SynthID. Finding the watermark is a strong clue that a file came from a participating tool. Not finding it says nothing about files from other tools, from tools that do not watermark, or from a real camera. Treat it as one clue next to the source, the original file and other checks; see how to show an image is not AI-generated.

TimeBinder does not detect AI content. If you made a file yourself, keep the original and the working files. A TimeBind records a SHA-256 fingerprint of the exact file and a time, anchored to Bitcoin. The fingerprint is made in your browser and the file is not uploaded. A TimeBind shows a file existed by a recorded time and has not changed since. It does not register copyright, is not legal advice and does not by itself show who made the file or whether it is genuine.

Make a TimeBind: first one free →

First Pooled TimeBind free, no card, then from US$5. An ASAP TimeBind is US$15.

Frequently asked questions

Is SynthID the same as C2PA Content Credentials?

No. SynthID is a watermark hidden in the content itself. Content Credentials attach signed provenance information to a file as metadata, which can be removed.

Can I check any AI image with SynthID?

Only for watermarks from Google and the partners it lists. An image from another tool may carry no SynthID watermark.

Does a missing SynthID watermark mean the image is real?

No. It only means no SynthID watermark was found.

Is SynthID only for images?

No. Google describes it for images, video, audio and text.

Sources

Google DeepMind, SynthID.
Google AI for Developers, SynthID Text: tools for watermarking and detecting LLM-generated text.
EUR-Lex, Regulation (EU) 2024/1689 (AI Act), Article 50.
TimeBinder, Does AI leave a watermark?.

Related guides and tools

Your file’s fingerprint, anchored to a global network secured by real electricity and cryptography.

Editing a file—including with AI—changes its fingerprint, not its existing network record.

How this works

Your file’s SHA-256 fingerprint, a short code worked out from its exact contents, is made in your browser, and the file itself stays on your device. A cryptographic commitment to that fingerprint, which may be combined with others, is then anchored in a block on the Bitcoin blockchain. Bitcoin is secured by proof of work: computers around the world spend real electricity to add each block, and changing an earlier block would mean redoing that work faster than the rest of the network. Until that block is confirmed, your proof is pending. Once the Bitcoin anchoring is confirmed, you can verify the proof against that recorded commitment. The confirmed record can then be checked; Pooled TimeBinds are checked through TimeBinder, while ASAP TimeBinds can also be checked independently. If you check the file later and its fingerprint still matches, that shows the file existed no later than the recorded time. The record doesn’t show who made the file, whether AI was used, or that its contents are true.

Your filestays on your device
FingerprintSHA-256
Network timestampon Bitcoin

Record a fingerprint of the file you rely on

Add a layer of protection against AI and human fraud and disputes: a dated, checkable record of the exact file you hold.

Open the app and sign in with email or Google.
Add your files. They stay on your computer.
Seal them and download your Proof of Time certificates.

Open TimeBinder →

First Pooled TimeBind free, no card, then from US$5. See pricing.