Google Launches Public SynthID Detector to Identify AI Watermarks

Google has opened public access to a standalone website called SynthID Detector that lets users check whether image, video or audio files carry its AI watermarking signals, broadening a verification tool that had previously been limited to early testers.

The company announced the launch in an Oct. 7 blog post, “We’re making it easier to identify AI-generated content globally,” written by Pushmeet Kohli, vice president for science and strategic initiatives at Google DeepMind. Google said the platform is available globally in English starting Wednesday and is designed to help users identify whether online content was created using Google AI or tools from industry partners. It said the detector can identify watermarking from Google and partners including OpenAI, NVIDIA, Kakao and, soon, Apple.

“We’re launching a standalone platform to help you easily identify whether online content was created using Google AI or tools from our industry partners,” Kohli wrote.

Google said the new site expands a set of existing verification features already built into Search, the Gemini app and Chrome. Those built-in tools now regularly handle more than 1 million requests a day, according to the company. Google also said, as of Oct. 7, it has watermarked more than 180 billion images and videos, along with 240,000 years of audio content.

“To date, we’ve watermarked over 180 billion images and videos, along with 240,000 years of audio content,” the company said in the post.

The launch marks a wider release of technology Google has been developing for several years. Google DeepMind introduced SynthID in 2023 as a watermarking system for AI-generated content. In May 2025, Google announced a SynthID Detector verification portal and initially offered access to journalists, media professionals and researchers through an early-access waitlist. Wednesday’s release is the broader public rollout of a standalone detection tool.

The practical use case is narrow but important. SynthID Detector checks for watermarking or provenance signals that participating AI systems add to content. It is not a universal AI detector, and a result showing no SynthID signal is not proof that a file was made by a person. The tool can only verify content generated by systems that actually apply the watermark.

The timing reflects a broader industry shift toward provenance tools meant to help identify synthetic media. In 2026, major AI companies, platforms and news organizations have increasingly focused on machine-readable signals such as invisible watermarking and C2PA, a technical standard for content metadata, as governments and regulators press for more transparency around AI-generated material. That push has also taken on added weight in Europe, where Article 50 transparency obligations under the EU AI Act took effect Aug. 2.

Google’s detector arrives as some of the biggest AI companies move closer together on those systems. OpenAI said in May 2026 that it would adopt SynthID watermarking for images as part of a broader provenance approach alongside C2PA metadata. Google is now positioning the public SynthID Detector as one place to check for those signals across its own tools and a growing set of partners.