As AI video models become increasingly powerful, YouTube is no longer relying solely on the integrity of content creators. The company announced on Wednesday that its internal systems will automatically apply special labels if they detect "significantly photorealistic AI" in a video. This was reported by Techcrunch.com reports .
YouTube is making its AI labels more prominent, allowing them to be quickly noticed in both long-form videos and YouTube Shorts. AI labels have been on the platform for over two years, and previously, creators were required via the Creator Studio tool to disclose content that could be mistaken for a real person, place, or event.
The company stated that its policy on AI labels has not changed, but the platform is now taking a more active role in content moderation. This step comes against the backdrop of Gemini Omni and new multimodal models that create high-quality videos, which were introduced at the Google I/O conference last week. Starting in May, YouTube began identifying AI products using new internal signals.
If creators forget to disclose that they used AI tools, YouTube will do it for them. Also, labels on videos created with YouTube's own AI tools, such as Veo or Dream Screen, cannot be removed. Furthermore, videos with C2PA standard metadata, which includes OpenAI, NVIDIA, and other companies, will also be permanently labeled.
This update follows the expansion of YouTube's deepfake detection system. Now, the platform offers the ability to scan and match faces not only for celebrities and politicians but for all users. While labels previously appeared only in the description, they will now be displayed directly on the video player itself.
Read “Zamin” on Telegram!
