YouTube announced on Wednesday a new feature called custom feeds, which enables users to create personalised video recommendation feeds using artificial intelligence. Users can type a description of the videos they wish to see into a prompt box, and Google's Gemini AI model will build a feed based on this request.
The company suggests examples such as a feed for video podcasts during a commute or one for specific moods like "relaxing commentary videos to unwind with." These AI-powered feeds can be described in detail, allowing users to specify content to feature, exclude, or prioritise. The resulting custom feed will be pinned to the top of the YouTube home page in its own tab.
YouTube believes custom feeds will assist users in exploring the platform's extensive video library. Emily Moxley, VP of Product Management for Viewer AI, noted that there are over 20 billion videos available. The new custom feeds will not replace YouTube's main recommendation feed but will offer an additional option for users.
Support for creating multiple custom feeds is scheduled to roll out across YouTube on web and mobile starting next month.