Google’s SynthID AI watermarking tech is being adopted by OpenAI, Nvidia, and more (arstechnica.com)

Google's SynthID AI watermarking technology is gaining traction as leading organizations like OpenAI and Nvidia begin adopting it. Initially showcased three years ago, SynthID has successfully tagged an impressive 100 billion images and videos and processed 60,000 years’ worth of audio. This significant expansion highlights the urgency to distinguish real content from AI-generated media in a landscape where digital authenticity is increasingly contested. Beyond SynthID, Google is integrating the C2PA standard into its products, which details the provenance of content through metadata, enhancing transparency. This functionality will soon apply to videos captured on Pixel 8, 9, and 10 devices and will also be incorporated into Google’s Gemini chatbot, Chrome, and Search. SynthID, which embeds a digital watermark directly into the pixels of images and audio waveforms, provides robust protection against removal, even after modifications like compression or cropping. This resilience is vital, as per Google DeepMind scientist Pushmeet Kohli, who emphasizes the importance of protecting against attempts to circumvent the watermarking. Overall, these advancements position Google at the forefront of efforts to ensure trustworthiness in AI-generated content.