Archive · 8 Oct 2026
Google Opens SynthID Detector Worldwide to Check AI Content
Google has opened a website that checks some pictures, videos and sounds for signs of AI.
It looks for an invisible mark called SynthID.
People upload a file, and the website checks whether the mark is there.
Some AI tools add this mark when they make or change media.
Google says the detector can check content from Google and several partner companies.
Support for Apple Intelligence images is expected later.
Google says the mark can survive some common changes to a file.
The website cannot find every kind of AI-made content.
It may also not tell whether AI made the whole file or only changed it.
Google has opened its SynthID Detector website to the public worldwide in English after earlier testing with media professionals and other users.
The free web tool checks uploaded images, videos and audio for invisible SynthID watermarks linked to supported AI tools.
Google says it supports content from Google, OpenAI, NVIDIA and Kakao, with Apple Intelligence image support expected soon.
Google says SynthID watermarks can remain detectable after common changes such as cropping, colour adjustments, frame-rate changes and lossy compression.
The detector only identifies SynthID-marked content and may not distinguish fully AI-generated material from content that was AI-edited.
- Who
- Google opened the SynthID Detector to users worldwide.
- What
- A web tool checks images, videos and audio for SynthID watermarks associated with supported AI tools.
- When
- Google announced worldwide availability; the articles do not give a specific calendar date.
- Where
- Worldwide on the web, currently in English.
- Why
- Google says the tool is intended to give people more context about content they encounter online.
Potential benefits
Limitations
Checking supported AI media
Potential benefits
The detector can flag supported images, videos and audio containing SynthID, including content linked to several named AI providers.
Limitations
It may not identify AI content that lacks a SynthID watermark or uses another watermarking system.
Understanding how content was made
Potential benefits
A SynthID result can indicate that a supported AI system created or modified media.
Limitations
The tool may not reveal whether the entire item was AI-generated or only edited with AI, nor which parts are synthetic.
- Cost
- Free to use
- Availability
- Worldwide on the web in English
- Supported media
- Images, videos and audio
- Named supported AI providers
- Google, OpenAI, NVIDIA and Kakao
- Apple support
- Apple Intelligence image support is expected soon
- Reported watermarking scale
- Google says SynthID has been used on more than 180 billion images and videos and around 240,000 years of audio
- Existing verification features
- Google says features across Search, Gemini and Chrome handle more than one million requests a day
- Limitations
- The detector looks for SynthID and may not flag media without that watermark or distinguish fully AI-created content from AI-edited content
Sources
Google Opens SynthID Detector To All: Lets Users Check Images, Video & Audio For AI Watermarks
Google Opens SynthID Detector to Everyone
Google Launches New SynthID Website To Check If Content Was Made By AI: How It Works
