Google has taken another step in the fight against misinformation generated by artificial intelligence. The company has opened the SynthID Detector to the general public, a tool capable of identifying whether an image, video, or audio was created with AI. Until now, access was limited to journalists and media professionals.
The announcement comes at a time when the proliferation of synthetic content makes it increasingly difficult to distinguish the real from the artificial. The measure responds to a growing demand: having accessible mechanisms to verify the origin of digital files circulating on the internet.
What is SynthID and how does it work
SynthID is the invisible watermarking technology that Google launched in 2023. It is a system that embeds an imperceptible signal into content generated by artificial intelligence, allowing its origin to be tracked later without altering the appearance of the file.
According to data provided by the company, this technology has already marked more than 180 billion images and videos, as well as 240,000 years of audio generated with AI tools.
Access and scope of the tool
- What it verifies: whether an image, video, or audio was generated with AI from Google or its technology partners.
- Who can use it: anyone, for free, via the website synthid.com.
- Available language: for now only in English, although with global reach.
- Integrated partners: in addition to Google's own models, the tool recognizes content generated by OpenAI, NVIDIA, and Kakao. Apple will be added soon.
The announcement was made on October 7, 2026, by Pushmeet Kohli, Vice President of Research and Strategic Initiatives at Google DeepMind.
One more piece within the Google ecosystem
The detector does not arrive in isolation. Google had already integrated similar verification functions into its search engine, the Gemini app, and the Chrome browser. According to the company, these three channels collectively receive more than one million daily queries related to the detection of AI-generated content.
With the opening of the SynthID Detector, Google seeks to consolidate a common verification infrastructure that encompasses both its own products and those of other technology companies that adopt its marking system.
Limitations to keep in mind
The tool can be useful for journalists, communications teams, or any user who receives an image or audio of dubious origin. However, its scope has significant limits that should be understood before fully trusting its results.
The detector only identifies the invisible mark of Google and the partners that have adopted its technology. This means that a negative result does not certify that the content is authentic or was created by a person: it only indicates that this specific mark was not found in the analyzed file.
Content generated with other AI tools that do not use SynthID could go unnoticed by the system. Furthermore, for the moment, the tool is only available in English, which limits its practical use for part of the Spanish-speaking public.
Synthetic content marking gains ground
The trend of invisibly identifying content generated by artificial intelligence is not exclusive to Google. Other companies in the sector have moved in similar directions in response to growing concern about digital misinformation.
This movement is also advancing in Europe, where various technology companies have announced measures to mark content generated by their artificial intelligence systems, in a regulatory context that increasingly demands greater transparency regarding the origin of digital content.
Frequently Asked Questions
What exactly is SynthID Detector?
It is a Google tool that allows you to check if an image, video, or audio was generated with artificial intelligence, by detecting an invisible watermark embedded by Google's models or those of its partners.
Who can use this tool?
Since October 7, 2026, anyone can access it for free via the website synthid.com. Previously, access was restricted to media professionals.
Does it detect content created with any AI?
No. It only identifies content generated with models from Google or partner companies that have adopted the SynthID technology, including OpenAI, NVIDIA, and Kakao. Apple will be added soon.
Does a negative result mean the content is real?
Not necessarily. A negative result only indicates that the invisible mark of Google or its partners was not detected, but it does not certify that the content was created by a person or with another AI tool.
In what languages is the tool available?
For the moment, SynthID Detector is only available in English, although its access is on a global scale.
Where else does Google use this marking technology?
Google already integrates similar verification functions into its search engine, the Gemini app, and the Chrome browser, which together receive more than one million daily queries.
