Summary
Google has opened SynthID Detector to the public, expanding access beyond the journalists, researchers and media professionals who previously used it. According to TreceBits, users can upload supported image, video or audio files at synthid.com and check for the imperceptible watermark embedded by compatible AI systems. The service can sometimes identify the portions of a file where the mark appears. Google says SynthID has been applied to more than 180 billion images and videos, plus audio equal to about 240,000 years. The English-only detector is not universal: if it finds no watermark, that does not establish that a file is authentic or human-made; it may simply come from an unsupported tool or model.
Hana’s perspective
For animation and post-production teams, this is useful infrastructure rather than a truth machine. A public check can help a producer, broadcaster or client confirm whether a delivered asset carries Google’s provenance signal, and the ability to localize the mark may be especially valuable in edited video. But a negative result proves very little: many generators do not use SynthID, and legitimate edits or mixed pipelines complicate provenance. I would treat the detector as one layer in a broader workflow that also preserves project files, model and prompt logs, rights documentation, and human sign-off. Its real value is making disclosure easier to inspect, not making synthetic media automatically trustworthy or untrustworthy.
Are you a writer with a great story? We'd love to hear from you: APPLY HERE