Progress in generative artificial intelligence enable people to create content in a completely new way – from text to high quality sound, images and movies. As these possibilities develop and become more accessible, questions about authenticity, context and verification arise.
Today we are announcing a synthesizer detector, a verification portal that quickly and efficiently identifies the content generated by AI made of Google AI. The portal provides the possibility of detecting various modalities in one place and provides the necessary transparency in the rapidly developing landscape of generative media. It can also emphasize which parts of the content are more often marked with water and synthesis.
When we launched a synthid-latest tool that embeds imperceptible aquatic signs and allows the identification of the content generated by the AI-Naszym goal was to provide a set of new technical solutions to help minimize disinformation and improper determination.
Synthid not only maintains the quality of the content, but works as a solid water sign that remains detectable, even when the content is available or undergoes a series of transformation. Although we originally focused only in the paintings generated by AI, we have expanded Synthid since then to include text, audio and video content generated by AI, including content generated by our twins, Imagen, Lyria and VEO in Google. Over 10 billion content has already been marked with a watermark with synthesis.
How the synthesizer detector works
After sending the image, audio path, video or a fragment of the text created using AI Google tools, the portal scans the media in search of a synthesis watermark. If the portal's watermark is detected, it will distinguish specific parts of the content that will most likely be significant.
In the case of audio, the portal indicates specific segments in which a synthesis watermark is detected, and for images it indicates areas in which the watermark is most likely.
