New YouTube AI labeling rules go live: Labels are more prominent, and will be applied even if creators don't disclose it

📄Full Article· Automatically extracted by trafilaturaGemini 翻譯1329 words

YouTube announces a major update: the platform is launching an internal system to automatically detect videos containing highly realistic AI-generated content and will automatically apply labels even if creators do not disclose it themselves. (Previous coverage: Tricking Google AI in 20 minutes: A single blog post can pollute the "only answer" seen by 2.5 billion people) (Background: The US aims to let AI grow wildly to "defeat China head-on," Trump's policy takes a sharp turn: leveling domestic state-level regulations) "Automatic AI detection is now live." The official announcement states that the new system will automatically label videos that "contain significant amounts of realistic AI usage." The judgment is based on internal detection signals from the platform; while specific technical details have not been fully disclosed, it is known to include analysis of the video's visual content. YouTube no longer relies solely on creator self-reporting. It announced, "For creators who believe they have been incorrectly labeled, YouTube provides an appeal process, allowing them to modify the disclosure status via YouTube Studio. However, the label will remain permanent and cannot be modified in two scenarios: First, content created using YouTube's own generative AI tools (including Veo and Dream Screen). Second, videos containing C2PA metadata that indicate they are fully AI-generated. C2PA (Coalition for Content Provenance and Authenticity) is an industry standard developed by organizations including Microsoft, Adobe, and the BBC, aimed at providing digital content with verifiable "provenance signatures." When a video's metadata is explicitly marked as AI-generated, the platform will not allow for modifications. In the past, YouTube's AI disclosure labels were hidden in the expanded section of the video description, a place most people never check. This update directly changes the placement of the labels. For long-form videos (non-Shorts), the AI label will move from the description box to directly below the player and above the description. This is where viewers' eyes naturally rest before or at the beginning of a video. YouTube Shorts are handled more directly: the AI label is overlaid on the video screen and cannot be ignored. Rene Ritchie, YouTube's Creator Liaison, stated in an official explanatory video: "The goal is to ensure viewers can grasp the context at a glance. If it looks real but is actually made by AI, viewers should know immediately." He also emphasized that the AI label itself does not affect a video's recommendation ranking, nor does it impact a creator's eligibility for monetization. YouTube's stance is: this is purely for information transparency, not a punitive mechanism. It is worth noting that the labeling system is not one-size-fits-all. For content that YouTube deems "unrealistic, animated, or slightly modified"—meaning AI-assisted creations that are clearly not realistic—the disclosure label will remain in the expanded description section rather than being moved to a prominent position. The trigger for prominent labels is locked to AI-generated content that "looks realistic." This label update is not an isolated policy adjustment but an extension of YouTube's recent series of AI governance actions. Earlier this month, YouTube opened its likeness detection program to all creators aged 18 and over. The purpose of this system is to help creators detect and manage "how AI depicts you on YouTube." In plain terms, if someone uses your face to generate a fake video, YouTube's system will help you identify it, and you can apply for its removal directly through YouTube Studio. YouTube has chosen the path of institutionalization rather than prohibition. It does not ban AI creation but demands visibility; it does not punish creators who use AI, but ensures viewers know what they are watching.

Data Status✓ Full text extractedRead Original (動區 BlockTempo)

🔍Historical Similar Events· Keyword + Asset Matching2 items

2026-05-19

Google releases the most powerful multimodal video model "Gemini Omni"! Supports conversational seamless photo editing, free on YouTube this week

Similarity 120%關鍵字 youtube同分類 zh

2026-05-17

YouTube launches AI deepfake detection tool: Creators 18+ can detect likenesses and request takedowns

Similarity 120%關鍵字 youtube同分類 zh

💡 Currently matching via keywords + symbols (MVP) · Will be upgraded to embedding semantic search later

Raw Information

ID:abaa398854

Source:動區 BlockTempo

Published:2026-05-28 01:28:37

Category:zh_news · Export Category zh

Symbols:Unspecified

Community Votes:+0 / −0 · ⭐ 0 Important · 💬 0 Comments