Identifying AI-Generated Content with SynthID and C2PA

An explanation of how to tell whether media is real or AI-generated, centered on Google DeepMind's SynthID watermarking and C2PA Content Credentials. It shows a Gemini app check and outlines partnerships and labeling efforts for AI-generated content.

■ Detecting AI-generated content
- Uploading media to Gemini app to ask if AI-generated
- Google DeepMind's SynthID as hidden watermark for AI images, videos, audio, and text

■ Scale and partnerships
- Over 100 billion images/videos and 60,000+ years of audio watermarked
- Partnerships with OpenAI, Nvidia, and Apple to apply SynthID to their content

■ C2PA Content Credentials and verification
- Credentials added to generative AI tools to indicate camera origin, AI editing, or full AI generation
- C2PA and SynthID together provide fuller context; verification can be tried by users

Suitable for viewers interested in AI transparency, content authenticity, and verification tools, it provides a concise overview of SynthID and C2PA and their combined role. Viewers are encouraged to try SynthID and C2PA verification to better understand the content they consume.