YouTube推出AI生成内容自动标签系统

New Prominence for AI Disclosures

  YouTube is launching a new automated system to detect and label photorealistic, AI-generated content to increase transparency for viewers. These labels will now appear in prominent positions directly below long-form videos and as permanent overlays on Shorts.

  YouTube is significantly tightening the transparency standards for artificial intelligence (AI) to tell the viewers if the content was manufactured using heavy AI generative tools. The platform has announced that it will now require the creators to disclose if their content has used realistic-looking AI to design images or videos. The platform has announced that it will identify the videos to look for signs of AI-generated content and subsequently will flag them.

  YouTube will deploy a new system that will automatically detect and label "photorealistic and meaningfully AI-altered" content, the company noted. The move marks a major shift for the platform, which previously relied almost exclusively on creators to voluntarily disclose the use of synthetic media.

  The update is rolling out this month and, as per the company, is designed to provide viewers with "context at a glance."

  While creators are still required to manually disclose AI usage during the upload process, YouTube's new internal detection signals will now act as a safeguard. If the platform's systems identify significant use of realistic AI that a creator failed to declare, the system will apply the disclosure label automatically.

New Prominence for AI Disclosures

  Apart from detecting the AI-generated content, YouTube is also changing where these labels appear to ensure they are impossible for viewers to miss. For long-form videos, the disclosure tag is moving from the buried "expanded description" section to a prominent position directly below the video player. For YouTube Shorts, the label will appear as a permanent visual overlay on the video itself.

  The platform, in a statement, has clarified that these heightened labels apply specifically to content that looks real but was generated or significantly altered by AI. Videos that are clearly animated, unrealistic, or only slightly edited will continue to carry their disclosures within the expanded description box rather than as a prominent, on-screen label.

What Are the Protections for Creator Control?

  YouTube has said that it has built in a mechanism for creators to dispute incorrect tags should they feel the system has glitched. If a creator believes their content was misidentified as AI-generated, they can update the disclosure status within YouTube Studio.

  However, this option is not universal. Labels will remain permanent and unremovable for content created using Google's native AI tools, such as Veo or Dream Screen, as well as for files that carry C2PA metadata, which serves as a verifiable, industry-standard digital watermark for AI provenance.

  For creators concerned about the potential impact on their livelihood, YouTube has offered strong reassurances. Rene Ritchie, YouTube's head of editorial and creator liaison, has emphasised that these labels are informational rather than punitive.

  The platform confirmed that the presence of an AI disclosure label will not affect a video's standing in recommendation algorithms, nor will it impact a creator's eligibility to earn ad revenue.

Why the Shift?

  The initiative comes amid mounting global pressure to curb the spread of AI-driven misinformation, particularly deepfakes and manipulated content involving public figures, elections, and financial topics.

  With the European Commission's AI Act transparency obligations set to take effect in August this year, the tech industry is under scrutiny to implement machine-readable markers. By moving toward automated detection, YouTube aims to strike a balance between providing necessary transparency to viewers and maintaining a reliable, creative ecosystem.