YouTube is launching a major update to its content labeling system: the platform will now automatically identify videos created or significantly altered using generative artificial intelligence. This move aims to increase transparency in an era where neural networks can produce photorealistic clips indistinguishable from real footage.
Previously, the responsibility for labeling AI-generated content rested solely with creators—they had to manually add the appropriate tag. Now, YouTube is enhancing this system with automated scanning: if algorithms detect signs of generative technology use and the creator hasn’t added a warning, the system will add one automatically.
This is especially relevant for videos featuring photorealistic effects—such as deepfakes or synthetic faces. Creators retain the right to contest a label if they believe it’s incorrect. An exception applies to videos created via official Google tools—such as Dream Screen or Veo—which will not be automatically labeled.
Labels will also remain on materials already containing a C2PA digital watermark—an international protocol for marking AI-generated content. This ensures that even when videos are transferred between platforms, their origin remains transparent.
For viewers, warnings will become more prominent: in regular videos, they’ll appear below the player; in Shorts, they’ll overlay the video itself to ensure they can’t be missed. The company emphasizes this responds to user demand for clearer distinctions between real and generated content.
The goal of this update isn’t to restrict creativity but to protect audiences from misinformation and manipulation. In a world where AI can generate any video, transparency becomes a key element of trust in the platform.