YouTube Rolls Out AI Video Auto-Detection System; Undisclosed Realistic Synthetic Content to Be Mandatorily Labeled

On June 1st, it was reported that YouTube is advancing an automated detection system specifically designed to identify generative AI video content on its platform, marking a shift from its previous reliance primarily on creator self-disclosure. According to this new mechanism, once the system identifies highly realistic videos that are not explicitly declared as AI-generated, it will automatically append a label such as 'AI-generated' or a similar prompt to the video interface. Creators can apply for the removal of such labels by updating their video information. However, if the platform's algorithm determines the content originates from AI tools like Google's Dream Screen or Veo, or detects internationally recognized anti-forgery metadata watermarks like C2PA, the relevant labels will remain enforced. This adjustment aims to address the ongoing risks posed by the proliferation of deepfakes, safeguarding users' right to know about video content while simultaneously strengthening the platform's review and management of synthetic media. [This article is sourced from Ebrun Go. An automated writing robot developed by Ebrun, delivering e-commerce industry intelligence to you first via algorithm. This dog is still very young; welcome to contact run@ebrun.com or leave a comment to help it grow.]