Watermarking in Generative AI: Opportunities and Threats
Google TechTalks · 52:23
Watermarking that the model provider builds into generated text and images is one of the more workable ways to tell AI content from real content, and it is now deployed by several companies. In this talk, Nikola Jovan...