Google's Invisible AI Watermark To Identify Generative Text And Video

At its I/O developer conference, Google introduced a range of new AI models and tools, including the expansion of its AI content watermarking and detection technology to encompass both text and video formats.

Google’s DeepMind CEO, Demis Hassabis, announced that the new upgraded SynthID watermark imprinting system will now mark video that was digitally generated as well as AI-generated text.

SynthID, unveiled last August, began as a tool designed to embed AI-generated images with imperceptible markers that humans can't see but the system can detect.

This method contrasts with other proposed watermarking standards, such as C2PA, which incorporates cryptographic metadata into AI-created content.

Google had also enabled SynthID to inject inaudible watermarks into AI-generated music that was made using DeepMind’s Lyria model.