YouTube continues to make strides in the field of AI-driven music creation. Following the launch of its feature last year that allowed creators to generate AI songs featuring the voices of renowned artists like Charlie Puth and Charli XCX through text prompts, the platform has recently announced an enhancement to its Dream Track toolkit with a song reshaping capability. This new feature enables creators to alter the style of songs through simple text descriptions.
This experimental functionality is currently available for testing by a select group of creators. Members of the test group can utilize the "reshape song style" option in chosen songs, inputting text descriptions of the desired changes. The system will then generate a 30-second audio clip for use in Shorts videos. YouTube explains that creators need only input their ideas for modifying the song's style or mood in the prompt box to receive a personalized track that retains the essence of the original vocals and lyrics while incorporating a new musical style.
To safeguard original rights, all reshaped audio clips will be clearly marked with the original song information and the use of AI technology in their adaptation, both on the Shorts video and audio pages. These features are powered by the Lyria music generation model developed by DeepMind. It's worth noting that the Dream Track toolkit, launched in November of last year, also includes a feature for creating songs through humming.
In balancing the interests of all parties in the music industry, YouTube announced as early as August of last year that it would compensate artists and copyright holders for the use of music in its AI features. The platform has also partnered with Universal Music Group (UMG) to develop a mechanism for distributing copyright royalties.
In the realm of music remixing, YouTube is not alone in its exploration. Former JioSaavn executive Gaurav Sharma is developing an app called Hook, which also aims to provide song remixing capabilities to support short video creation.
These initiatives demonstrate that AI music creation is moving towards a more regulated and mature phase of development, offering creators more creative possibilities while actively seeking solutions to protect original rights.