Google announced that it has added the ability to produce music to the Gemini application. The company uses DeepMind’s Lyria 3 music production model to provide this feature. This feature is currely in beta and is available for free.
To use this possibility, the user has to describe the song he was to make, and then the application produces a piece of music along with the lyrics. For example, a Gemini could be asked to create a humorous and light-hearted Arrandb piece about a sock finding its mate; In this case, the program will produce a 30-second piece with a cover image created by Nano Banana.
Google has announced that users can even upload a photo or video and this AI-based tool will create a song to match the mood of the media file.
The company also said that the Lyria 3 is an improveme over the previous generation models and can produce more realistic and complex pieces of music. Users can also change and corol other elemes such as style, type of singing and tempo.
Simultaneously with the release of Lyria 3 in the Gemini application, Google also provides this model to coe producers through the Dream Track feature on YouTube. Dream Track is a tool that helps creators create AI-generated music tracks. This option was previously only available to YouTube producers in the US, but with this release, Google is expanding Dream Track’s reach globally.
Google has stated that it is not possible to directly imitate an artist, but if an artist’s name is eered in the command, Gemini will produce a piece with a similar style or mood. It remains to be seen whether this capability will make it easier for others to produce, analyze and ideify a particular artist’s musical style.
In a blog post, Google stated that Lyria 3’s music production capabilities are designed for creative and original expression, not imitation of existing artists. If a specific artist is meioned in a user’s order, Gemini will consider that as the overall creative inspiration and produce a piece that is similar in style or mood. The company has also announced that filters are in place to check outputs against existing works.
Google meioned that all songs produced with the Lyria 3 model will have a SyhID watermark so that the coe produced by artificial ielligence can be ideified. The company added that inside Gemini, it will also add capabilities to ideify music produced by artificial ielligence using SyhID. Users can upload pieces of music and ask Gemini if the piece was produced by AI.
The music production feature is available to all Gemini users 18+ worldwide and supports English, German, Spanish, French, Hindi, Japanese, Korean and Portuguese.
Music produced by artificial ielligence has evoked mixed reactions among artists and listeners. On the one hand, companies like YouTube and Spotify are embracing AI and signing deals with music labels to monetize AI-produced music. On the other hand, companies developing artificial ielligence models and tools have faced lawsuits from the music industry regarding the copyright of educational data. Platforms such as Deezer have also released tools to flag AI-produced music and preve the rise of fraudule streams of this type of music.
We have produced the following file for free with this model:
