Previously available only to journalists and researchers, the tool recognises the SynthID watermark used by Google, as well as by OpenAI, Nvidia and Kakao.
Google has made the SynthID Detector site available to everyone, which allows you to submit a file to find out if it has been produced by an AI model. The tool was presented at last year's Google I/O conference, but until now only certain journalists, media professionals and researchers have been able to use it. It accepts a wide range of formats: a dozen for images (including JPG, PNG, WEBP, AVIF or HEIC), MP4, MOV and WEBM for video, and WAV, MP3, FLAC or AAC, among others, for audio.
The detection is based on SynthID, the invisible watermark launched by Google in 2023 and integrated into everything generated by its Nano Banana, Veo and Lyria models as well as tools such as Gemini, Flow or Vids. OpenAI, Nvidia and Kakao have adopted the same branding, and Apple is reportedly preparing to support it. The verification is already available in the Gemini app and in Chrome, where Google says it receives about one million requests per day.
The scope of the tool remains limited to watermarked content: Microsoft and Meta use their own standards, and these tagging systems regularly fail, including on content from their own designers' templates.
Read the full article on the original source
Go to the source →