YouTube begins automatically labeling videos with AI.

YouTube has just announced a major change in how it manages artificial intelligence (AI) content on its platform, no longer relying solely on creators to declare their videos as AI-generated. Instead, YouTube's system will proactively detect and automatically label videos that use AI to create highly authentic content.

AI labels will be added to AI videos . Photo: YouTube

According to YouTube, new internal systems have been deployed since May to identify videos that use “significant optical AI.” When content created or edited with AI but not declared is detected, the platform will automatically add an AI label to the video.

In addition to expanding detection capabilities, YouTube has also changed how AI labels are displayed to make them more recognizable. Previously, this information typically only appeared in the extended description, except for videos with sensitive topics such as health or breaking news, which were displayed directly on the video. With the new change, AI labels will appear right below the video player for longer content and will be inserted directly into short YouTube Shorts videos.

YouTube says its policy regarding AI labeling remains unchanged, but the platform will play a more proactive role in monitoring. This move comes amid the rise of increasingly powerful AI video creation models capable of producing content that is difficult to distinguish from reality.

Especially after Google recently launched Gemini Omni, a new line of multimodal AI models at the Google I/O developer conference, capable of creating high-quality videos that reflect an understanding of physics, culture, history, and science .

YouTube's AI labeling policy has been in place for over two years, after the platform required creators to declare if videos contained AI content that could mislead viewers about real people, places, or events. However, content that is clearly animated, fantastical, or unrealistic—such as images of a unicorn running in a fictional world—is not required to be labeled as AI.

Nevertheless, YouTube emphasizes that the responsibility for reporting still rests with the content creator. If a video is incorrectly identified by the system, the channel owner can update the AI ​​status. However, labels related to content created using YouTube's own AI tools, such as Veo or Dream Screen, cannot be removed.

In addition, videos containing C2PA metadata—a standard that helps verify the origin of AI content—will also be permanently labeled to indicate that the video was created entirely using AI.

One notable point is that the AI ​​labeling will not affect monetization capabilities or how the system recommends videos. YouTube asserts that the goal of this change is to help users easily identify content that has been edited or created using AI to resemble real events.

Alongside strengthening its control over AI content, YouTube is also continuing to expand the application of AI on its platform with features such as interactive search, Ask YouTube, automatic playlist creation for YouTube Music, AI-powered video summarization, and many new content creation tools.

The automatic AI labeling marks a new step for YouTube in its efforts to balance AI technology development and ensure transparency, especially as AI videos are becoming increasingly difficult to distinguish from real events.

LE PHI (According to Techcrunch)

Source: https://baocantho.com.vn/youtube-bat-dau-tu-dong-gan-nhan-video-ai-a205754.html