YouTube is launching an experimental feature which lets viewers leave "Notes" under videos for further context and information, the company told us exclusively. If the feature sounds familiar, it is because it follows the same approach as Community Notes on X.

According to the owner of Google, its feature is intended for such uses as clarifying when a song is intended to be a parody or to alert the viewers that footage has been incorrectly dated.

The launch comes at a critical U.S. election year. While the problem of misinformation had been somewhat of a factor during the 2020 presidential election, the 2024 election promises to be even more dangerous with the newly emerging generative AI. With this new addition to YouTube, it would seem that the only course of action will be to restrict the circulation of misinformation on YouTube.

It will debut in mobile in the U.S. with the language being English.

YouTube recognizes that errors will indeed happen during this testing period because it anticipates that some notes will not work well for the video, or worse, contain an error in a note. During the testing period and still today, it is learning from the testing period and actively incorporating feedback of viewers and creators toward the quality of their notes.

The test phase includes inviting a few users to jot down some notes. Eligible users will have an active channel on YouTube in good standing.

Throughout the following weeks and months, viewers in the U.S. will begin to see annotations on videos. In the first pilot, third-party raters will rate the helpfulness of and accuracy of annotations. YouTube will use these responses to train their systems.

If assessment librarians or other third-party evaluators believe the notes are useful, they will appear in a list below a video. Later, audiences will be asked to vote on whether the note is "helpful," "somewhat helpful" or "not helpful." Then, there will be another question that expresses their reason for believing it helpful or unhelpful. For instance, an audience member might say that a note was helpful because it cited good sources or was neutral in its attitude.

YouTube will then use an algorithm for deciding which notes get published depending on how the notes have done based on ratings. The algorithm will find those that are useful to most people. Suppose most people who had rated the other notes otherwise now found a particular note to be helpful. Then, YouTube is likely to publish that note under the video.

The company says the system will continue to get better as more notes are submitted and rated across different types of videos on the platform. As it improves the feature, YouTube will also decide whether or not it makes sense to officially roll it out.