# YouTube将尝试从本月开始自动标记AI视频

- 来源：The Decoder：AI News（RSS）
- 作者：Matthias Bastian
- 发布时间：2026-05-28 00:54
- AIHOT 分数：61
- AIHOT 链接：https://aihot.virxact.com/items/cmpobdifh04mislv46jqi7oqq
- 原文链接：https://the-decoder.com/youtube-will-try-to-automatically-flag-ai-videos-starting-this-month

## AI 摘要

YouTube将上线一套新的AI内容标签系统，旨在使标签位置更显眼：长视频的标签将显示在播放器下方，Shorts的标签将作为叠加层显示。从2026年5月开始，该平台将启动自动检测系统，即使创作者未主动披露，也会对AI生成内容进行标记。该系统对视频的推荐和变现没有影响。

## 正文

YouTube will try to automatically flag AI videos starting this month

YouTube is overhauling its AI labeling system for videos. The company says recommendations and monetization won't be affected.

Starting now, labels for photorealistic or heavily AI-altered content will show up in more visible spots: right below the player for long videos and as an overlay on Shorts. Starting May 2026, YouTube will also roll out automatic detection of AI-generated content. If creators don't disclose their use of AI but the system picks up heavy use of photorealistic AI, a label gets applied automatically.

Creators can appeal, but some labels are permanent

Creators who think their content was wrongly flagged can update the disclosure in YouTube Studio. But labels are permanent for content made with YouTube's own AI tools like Veo or Dream Screen, and for content with C2PA metadata confirming full AI generation. So it seems like YouTube can enforce labels on its own tools, but for third-party AI content, it still depends on creators being honest or its detection system catching it.

YouTube says the labels won't affect a video's recommendations or monetization. The platform has been labeling AI content based on creator self-reporting since 2024. The company is also opening up its deepfake detection tool, "Likeness Detection," to all creators aged 18 and older.

Google is responding to a growing problem. Late last year, an investigation showed that YouTube is already drowning in low-quality AI content, some of it with political overtones. The flood of AI slop is hitting other platforms too. LinkedIn, for example, is trying to crack down harder on generated text.

AI News Without the Hype – Curated by Humans