Google has introduced a new online tool called SynthID Detector, designed to help users identify whether an image, video, audio clip, or piece of text was created using its own AI technology. The tool, unveiled during the company’s I/O 2025 event, is currently available to a limited group of beta testers.

Figure 1. Google.

core, SynthID Detector functions like a reverse fingerprint scanner. Users upload content, and the system scans for subtle, invisible watermarks embedded by Google’s AI models. If these digital markers are found, the tool flags the content—either in whole or in part—as AI-generated. Figure 1. Shows Google.

The launch comes at a crucial moment. AI-generated content and misinformation are rapidly spreading across social media and search engines. Deepfake videos, in particular, have surged, increasing more than fivefold since 2019. In fact, some of the most viral Facebook posts in the U.S. last fall were reportedly made using AI.

That said, SynthID Detector has its limitations. It can only identify media marked with Google’s proprietary watermarking system—so content generated by AI from companies like OpenAI, Meta, or Microsoft won’t be detected. Even Google's own system isn’t foolproof [1]: the company acknowledges that AI-generated text may go unnoticed if it’s significantly altered or reformatted.

Despite these shortcomings, Google emphasizes the scale of adoption. Since SynthID was first introduced in 2023, more than 10 billion pieces of content have been watermarked, according to internal figures.

Whether this will be enough to stem the tide of synthetic media remains uncertain. But for now, Google sees increased transparency as a meaningful step forward.

Reference:

1. https://www.digitalinformationworld.com/2025/04/google-maps-user-spots-mysterious.html

Cite this article:

Keerthana S (2025), Google Unveils New Tool to Detect Its Own, AnaTechMaz, pp.327