Google opens SynthID Detector site for checking AI media watermarks
A dedicated site extends watermark checks beyond an earlier limited test, but a negative result cannot prove that media was made without AI.
A wider route to watermark checks
Google has opened a dedicated SynthID Detector website beyond the journalists, media workers and researchers who could test it earlier, TechCrunch reported on October 7. The live portal presents Google’s service terms before use. According to TechCrunch, the site accepts common image, video and audio file formats and checks for the invisible SynthID marks added during supported AI generation. Google uses those marks in media from products including Nano Banana, Veo and Lyria. The new access route is distinct from the older SynthID watermarking technology and from September’s SynthID Bio work on protein designs.
What a result can establish
A detected mark can help trace supported media to a participating generation system, but the absence of a mark is not evidence that a file is human-made. Other systems may use different marks or none, and edits can complicate verification. TechCrunch also notes that watermark detectors have missed some altered content. The reviewed sources establish that the public portal exists and describe its supported media categories; they do not independently measure its detection accuracy or test every format. Users should treat a result as one provenance signal, not a universal ruling on whether a file was made with AI.