YouTube Music to Label Songs Made With AI, Visible on the Playback Screen
IT DAILY · · 2 views
✦ AI Summary
YouTube will introduce labels so users can tell right in YouTube Music when AI played a significant role in producing a song.
The labels will be added using both music supplier information and the platform's detection tools, and can be checked on the playback screen.
Songs that used generative AI are divided into three categories: Fully Gen AI, Partly Gen AI, and No Gen AI, and the same disclosure standard will also apply to general video creators.
YouTube will introduce labels in YouTube Music so users can tell right on the playback screen when AI played a significant role in a song's production. In an official blog post on the 14th local time, YouTube said it would roll out YouTube Music AI music labels gradually. The move is intended to help users understand how music is made.
The labels will mainly apply to songs produced with generative AI. This includes cases in which generative models were used for composition and for vocal generation. YouTube said the change reflects user requests to know whether they are listening to AI-generated music.
The labels will be added using a combination of information provided by music suppliers and the platform's own detection tools. Through this, YouTube plans to indicate when AI was used.
The labels will appear on the YouTube Music playback screen. The key point is that users will be able to confirm AI use directly on the listening screen. Songs in which AI played a significant role in production will carry a label on the player page.
If a user selects the label, additional information about the production method will be provided. The display is also expected to expand to the main YouTube app in the future.
This guidance concerns labeling whether generative AI was used during music listening. The scope of the official announcement is limited to indicating whether AI was involved in production. It does not include song-by-song disclosure of which AI model was used, nor does it include disclosure of training data sources.
Detailed guidance for music suppliers is organized into three categories of generative AI use. Category 1, "Fully Gen AI," refers to cases in which the entire content was produced with generative AI. Category 2, "Partly Gen AI," refers to cases in which generative AI was used for part of the work. Category 3, "No Gen AI," refers to cases with no content made using generative AI.
YouTube also gave examples for each category. If someone entered text prompts into an AI music generator and then downloaded the finished song, or if composition, instrument performance, and vocals were all produced with generative AI, that would fall under "Fully Gen AI." If AI was used to generate bass or string parts and a person separately recorded the vocals or instruments, or if generative AI was used to create lyrics or melody, that would fall under "Partly Gen AI."
Cases in which a person's singing was corrected using general pitch-correction software such as Auto-Tune, or in which an automated AI mastering tool adjusted the sound balance, are both presented as "No Gen AI." This is because the standard does not classify content simply because AI technology was used; instead, it distinguishes whether generative AI created creative elements.
The AI label itself is not a reason for revenue restrictions, but repeated failure to disclose will trigger separate action. YouTube determines what should be labeled by combining generative AI use information submitted by music partners with additional signals.
If a provider does not submit the relevant information, the use status is treated by default as unknown. However, if other signals make it possible to judge, the content may be designated as fully or partly AI-generated, and a lack of submission alone does not mean the song will be treated as not using AI.
YouTube applies disclosure standards for AI-generated music to general video creators as well. Examples of content subject to disclosure under YouTube's generative AI content guidelines include music made with AI. Creators can disclose this through the "AI use" item in the YouTube Studio upload process. It has also enabled labels to be automatically attached when internal systems detect AI-generated or AI-altered content.
However, disclosure of AI use is separate from penalties for policy violations. YouTube said disclosure of AI content alone does not limit the audience. It also said disclosure of AI content alone does not affect monetization eligibility. The issue arises when disclosure obligations are repeatedly not met. Possible measures for repeated nondisclosure include applying labels manually, removing content, and suspending participation in the YouTube Partner Program.
YouTube has also put in place a correction process for false detection results. Under the general content guidelines, creators can usually edit AI disclosure information in YouTube Studio. There are, however, restrictions on some edits. Changes are limited for content made with YouTube's own AI tools, content that includes C2PA metadata containing content source and production history, and content that was labeled after manual review.
Source: IT DAILY · Lee Jae-young
Original: https://www.itdaily.kr/news/articleView.html?idxno=242198
References
This article was produced with the help of an automated content generation algorithm.
Source: IT DAILY
View originalThis article was summarized and organized by BizCrush based on the original article from IT DAILY. For exact quotations and full details, please refer to the original article.