Google Releases SynthID Detector Tool to the Public

The SynthID Detector is a public web-based verification tool developed by Google that identifies artificial intelligence watermarks embedded in digital content. Falling into the category of AI provenance and detection software, the application allows users to upload media directly to its website to verify whether content contains SynthID invisible watermarking technology, addressing the growing challenge of identifying AI-generated or AI-edited imagery and video.

Key Facts

Attribute Value

Developer Google

Tool Name SynthID Detector

Technology Type Invisible AI watermarking detector

Supported Models and Partners Google Gemini, OpenAI, NVIDIA, Kakao, and soon, Apple

Unsupported Platforms Grok, Midjourney

Access Model Public web availability via official URL

Pricing Unknown (not specified in source material)

Release Date Unknown (specific launch date not specified in source material)

Google has made its SynthID Detector publicly accessible on the web to identify invisible watermarks embedded in synthetic media generated across supported partner platforms.

How Does the SynthID Detector Work?

The SynthID Detector works by analyzing uploaded image or video files specifically for Google’s proprietary invisible SynthID watermarks embedded during generation. When a user submits content to the website, the scanner evaluates the underlying digital structure to confirm whether the specific watermark signature is present across supported ecosystem partners.

According to Google’s announcement, the verification tool is strictly limited to media originating from services utilizing the specific watermarking standard. If the tool analyzes an upload and confirms a match, it displays the notification that “SynthID was detected.” This response guarantees the media was AI-generated or modified using compatible generation systems, even if the software was only utilized for light editing.

“This is not a general AI detector. We can only detect media from the above companies, who have adopted SynthID technology.” Google Announcement Page and FAQ Section

A positive “SynthID was detected” result guarantees that an uploaded image or video was generated or modified using tools integrated with SynthID technology.

What Are the Limitations of SynthID Media Identification?

The SynthID Detector cannot identify all synthetic media because it operates exclusively as a signature verifier rather than a universal AI detection platform. It only recognizes content produced by Google Gemini and participating partners, leaving media generated by widely used unpartnered platforms entirely undetected by the scanning interface.

Consequently, receiving a result stating “SynthID was not detected” does not guarantee that the submitted asset is authentic or free from artificial intelligence modification. Popular AI systems such as Midjourney and Grok do not deploy SynthID watermarks, meaning their outputs will not be flagged by the scanner. Furthermore, the exact accuracy rates, file size thresholds, and processing speeds remain unknown based on available documentation.

The SynthID Detector tool does not guarantee that unflagged media is free of AI manipulation because it cannot scan outputs from non-participating generators like Grok and Midjourney.

How It Compares

The SynthID Detector distinguishes itself from hypothetical universal detectors by serving as a specialized ecosystem tool rather than a generalized scanning engine. Its effectiveness depends entirely on industry adoption of Google’s watermarking standard among competing artificial intelligence model developers and platforms.

Platform / Model SynthID Detection Status

Google Gemini Supported (Native)

OpenAI Supported Partner

NVIDIA Supported Partner

Kakao Supported Partner

Apple Pending (“and soon, Apple”)

Grok Not Supported

Midjourney Not Supported

Platform compatibility for the SynthID Detector is restricted to Google Gemini alongside partner networks including OpenAI, NVIDIA, Kakao, and upcoming integration for Apple.

Common Questions

What does a “SynthID was detected” result mean?

A “SynthID was detected” result confirms with certainty that the submitted image or clip was generated or edited using artificial intelligence tools from Google or participating partners that embed invisible SynthID watermarks.

Does a negative result prove an image is not AI-generated?

No, a negative result does not prove content is authentic. The tool only scans for SynthID watermarks, meaning AI-generated media created by platforms lacking SynthID integration will return an undetected notification.

Can the tool detect media created by Midjourney or Grok?

No, the tool cannot detect media created by Midjourney or Grok. Because neither service is listed among the partners adopting SynthID technology, their synthetic content bypasses detection entirely.

Sources and Methodology

This article synthesizes technical reporting and official disclosures sourced directly from Lowyat.NET (“Google Releases SynthID Detector To The Public”), referencing official communications from Google DeepMind and the official SynthID portal. All reported entities, quotes, and platform partner listings are directly transcribed from the source context. No data conversions or currency translations were required. This article was last updated on October 8, 2026.