Back in May, YouTube introduced video chapters, a way of breaking down videos into separate, easy-jump-to segments. It's similar to the way a film is broken down into scenes on a DVD, allowing you to skip straight to the part you want. Sadly, there aren't many creators using chapters yet, mainly because it takes time to set them during the editing process. To help remedy this, YouTube has decided to use machine learning to assign chapters automatically, and testing is now underway.
Machine learning is being used to analyze what's happening in a scene, especially if there's any text being shown. From there, YouTube will auto-generate chapter names and implement them in the video, without any direct input from the creator required. Thankfully, this isn't going to be forced on anyone, and creators can opt out if they don't like what they see. YouTube also says that this will be tested on a small collection of videos, but I assume it will spread to more if the testing is ultimately successful.
So long as this ends up works reliably, it'll be a great addition to YouTube's feature set. We, the viewers, will be able to skip to the part of a video we want to watch, and the content creator won't have to lift a finger. Let's hope it all goes well.
- Google Support