Google has officially announced the integration of a music-generation feature into its Gemini application, leveraging DeepMind's advanced Lyria 3 music-generation model. This new capability, currently in beta, allows users to describe their desired song, after which the app will generate a 30-second track complete with lyrics and even cover art, often created by Nano Banana. For example, a user could request a "comical R&B slow jam about a sock finding its match," and Gemini would deliver a corresponding musical piece. Furthermore, the AI-powered tool can generate a song to match the mood of an uploaded photo or video.

The Lyria 3 model represents a significant improvement over its predecessors, capable of creating more realistic and complex music tracks. Users gain enhanced control over various elements, including style, vocals, and tempo, allowing for greater customization of their generated music. Beyond the Gemini app, Google is also extending Lyria 3's capabilities to YouTube creators through the Dream Track feature. Previously restricted to U.S. creators, Dream Track is now being expanded globally, providing a tool for more creators to produce AI-generated music.

Google has clarified its stance on artist mimicry, stating that the music generation with Lyria 3 is designed for original expression rather than direct imitation of existing artists. If a prompt includes a specific artist's name, Gemini will interpret it as broad creative inspiration, generating a track that shares a similar style or mood without mimicking the artist outright. The company also employs filters to ensure outputs are checked against existing content.

To maintain transparency and identify AI-generated content, all songs created with the Lyria 3 model will incorporate a SynthID watermark. Google is also adding features within Gemini that enable users to upload tracks and ascertain if they are AI-generated using SynthID. This music generation feature is being rolled out globally to all Gemini users aged 18 and above, with support for multiple languages including English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese.

The advent of AI-generated music has elicited a range of reactions across the music industry. On one hand, major platforms like YouTube and Spotify are embracing AI, signing contracts with music labels to facilitate the monetization of AI-generated content. On the other hand, AI model and tooling companies are encountering legal challenges from the music industry, particularly concerning copyright infringement related to their training material. In response, some platforms, such as Deezer, have introduced tools to mark AI-generated music, aiming to prevent fraudulent streams and uphold content integrity.