Google is using YouTube videos to train its AI video generator

Core Insights - Google is utilizing its extensive library of 20 billion YouTube videos to train its AI models, including Gemini and the Veo 3 video and audio generator [2][5] - The company confirmed that it uses a subset of videos for training while adhering to agreements with creators and media companies [2][3] - Experts warn that this practice could lead to an intellectual property crisis for content creators and media organizations [3][4] Company Practices - YouTube has stated that it has always used its content to enhance product offerings and is committed to providing protections for creators in the AI era [3][6] - The platform does not disclose how many videos are used for AI training, but even 1% of its catalog would equate to 2.3 billion minutes of content, significantly more than competitors [5][6] - Users who upload content to YouTube cannot opt out of having their videos used for AI training [6] Industry Implications - There is a lack of awareness among creators and media professionals regarding the use of their content for training Google's AI models [4][7] - Concerns have been raised about the fairness of using creators' work to develop synthetic versions of their content without proper acknowledgment or compensation [7]