YouTube Videos: A Tool for Responsible AI Training at Google

YouTube is utilizing its videos to train artificial intelligence models, including Gemini and Veo 3, a text-to-video generator from its parent company, Google. This initiative has been in progress and was highlighted in a September blog post but gained attention following a CNBC report detailing that content creators cannot opt out of having their videos used for this purpose. YouTube claims this practice aims to enhance the experience for both viewers and creators, promoting new generative AI features like auto dubbing and refining recommendation systems, though specifics on responsible use remain vague. While YouTube acknowledges the need for safeguards and is investing in protections for creators, the exact number of videos utilized for AI training out of the over 20 billion available is unclear. Google recently introduced its AI filmmaking tool Flow to assist directors and revealed its new investment in Promise, a L.A.-based film studio.

Source link