Technology 8 sources · today Latest coverage 7 Oct 2026, 7:44 pm UTC

Google launches SynthID Detector globally to check for AI-generated media

The free website lets people check images, videos and audio for hidden AI watermarks, while expanding partnerships could broaden the range of content it recognises.

By Hushread Stories, written with AI from 8 outlets · First published 7 Oct 2026

In brief

  1. Google has launched SynthID Detector globally, opening its AI media checking website to everyone.
  2. The free service checks images, videos and audio files for a hidden watermark called SynthID.
  3. Google's Gemini models label their AI content with SynthID, giving the detector a signal to identify.
  4. Google announced expanded detector partnerships with OpenAI, NVIDIA and Kakao alongside the public website launch.
  5. Apple support is described as forthcoming, and the website does not identify every piece of AI content.
Google launches SynthID Detector globally to check for AI-generated media
Source: DEV Community

Timeline · 4 moments

4 moments Open the full timeline →

Google expands partnerships with OpenAI, NVIDIA and Kakao

DEV Community ↗

Gemini content carries hidden SynthID labels for detection

Ars Technica Tech ↗

Public website supports image, video and audio checks

TechCrunch ↗

Google announces standalone SynthID platform for identifying AI content

The Keyword ↗

How it started

SynthID Detector rests on a hidden label, rather than a promise to recognise every AI-made image by its appearance. The Next Web reported that Google's AI tools add the SynthID watermark to their output, and the website checks whether that signal is present.

According to Ars Technica Tech, every piece of AI content from Google's Gemini models carries a hidden SynthID label. The new website makes checking for that label available to a wider public.

How it unfolded

On October 7, 2026, Google announced a standalone platform for identifying content created using Google AI or tools from industry partners, according to its announcement on The Keyword.

Coverage published that day described a public service covering several media formats. TechCrunch reported that anyone could check an image, video or audio clip. The Next Web described the website as free and open to everyone.

The rollout also expanded the companies involved. DEV Community reported partnerships with OpenAI, NVIDIA and Kakao, with Apple identified as a forthcoming participant.

There is a difference in how the coverage describes Apple's availability. Engadget said the detector would flag content from Apple alongside other companies, while 9to5Google explicitly described support for Apple's models as coming "soon". The coverage does not establish that Apple support is already live.

Where it stands

SynthID Detector is now available globally. Digital Trends reported that anyone can use it to check images, videos and audio for AI-generated content carrying its watermark.

That last qualification matters. This is a watermark check, not a universal test of whether media is authentic. The Next Web describes detection of the hidden SynthID signal, while tech cautions that the tool does not see everything. A check that finds no watermark should not be treated as proof that a file was made without AI.

What to watch

Apple participation is the clearest pending development. 9to5Google says support is coming soon, and DEV Community calls Apple a forthcoming participant. Neither provided summary gives a launch date.

The other open question is the breadth of partner support available now. Google's announcement on The Keyword describes identification across its own tools and industry partners, but the supplied coverage does not specify which models or outputs from each partner the website currently recognises.

Spotted an error? Email [email protected] and we'll fix it.

More in Technology

All →