Google expands SynthID for labeling AI content
Google notes that as generative AI evolves, it is becoming increasingly important to distinguish content created by neural networks from ordinary content. To address this, Google DeepMind launched SynthID back in 2023, a technology that adds a hidden digital watermark to AI-generated content.
Initially, SynthID worked with images, but it now also supports video, audio, and text. According to Google, the technology has already watermarked more than 100 billion images and videos, as well as an amount of audio equivalent to 60,000 years of sound.
Content verification through SynthID is now available in Google Search, Gemini in Chrome, and the Gemini app. Google is also implementing C2PA Content Credentials so users can see not only the watermark, but also the origin of an image or video and its edit history.
Why it matters
- —It is becoming easier for users to understand whether content was created or modified using AI.
- —SynthID is already used in major Google products, including Search and Gemini.
- —Opening the technology for text and collaborating with other companies could help make AI content labeling more widespread.
Key facts
- SynthID was launched by Google DeepMind in 2023 for invisible labeling of AI content.
- The technology now supports images, video, audio, and text.
- Google reports more than 100 billion labeled images and videos and 60,000 years of audio.
- SynthID verification is available in Google Search, Gemini in Chrome, and the Gemini app.
- Google is also implementing C2PA Content Credentials in generative AI tools.
The full text is in the original source. Here we provide a brief summary and key facts.