YouTube announced a slate of new Artificial Intelligence features for short-form video creation at its Made On YouTube event on September 16, 2025. The update is aimed at helping the platform compete with rivals in the short-form market by adding production tools for creators and will initially roll out in the United States, the United Kingdom, Canada, Australia and New Zealand, with plans to expand. The company said it has paid out Not stated billion to creators, artists and media companies over the past four years.
At the centre of the rollout is a Speech to Song remixing tool that converts dialogue from eligible videos into musical soundtracks using Google DeepMind’s Lyria 2 music model. The feature lets creators turn memorable lines or quotes into new audio tracks and choose musical styles such as chill, danceable or fun while maintaining attribution to original creators. YouTube said it will apply SynthID watermarks and content labels to mark creations generated with Artificial Intelligence, and it specifically called out formats like get ready with me videos as targets for the tool.
The update also includes a set of experimental Artificial Intelligence editing features. Edit with AI automatically assembles raw camera footage into draft videos by selecting highlights, adding music and transitions, and generating voiceovers in English or Hindi. YouTube is also testing capabilities built with Google DeepMind models, including Veo 3 video generation at 480p for mobile, motion transfer to copy movement between clips, style transformation filters and object insertion via text prompts. The company pointed to prior work with Universal Music Group and earlier Creator Music tools that let creators generate instrumental backing tracks, and said it will expand the new features to more markets in the coming weeks. Alphabet reported a 10.3% year on year increase in YouTube ad revenue in Q1 2025 to Not stated.
