Back to all articles
•Technology

Google Introduces AI-Powered Music Generation Feature in Gemini App with Global Expansion

View original source

Google has announced the addition of a music-generation feature to its Gemini app, leveraging DeepMind’s Lyria 3 model. This feature is aimed at creating music tracks and lyrics based on user descriptions. Here are the key points:

  • Users can describe the desired genre or theme, such as a “comical R&B slow jam,” and the app will generate a 30-second track complete with cover art by Nano Banana.
  • Users can further customize musical elements by adjusting style, vocals, and tempo, and even match songs to photos or videos for mood compatibility.
  • This deployment is in beta and is being rolled out to all users above age 18 globally, with support for multiple languages, including English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese.
  • Google is extending this functionality to YouTube creators worldwide via the Dream Track tool, previously limited to U.S. users only.
  • The company emphasized that while the AI can generate music inspired by specific artists, it is not designed to exactly mimic or replicate existing artists’ work. This is part of a broader effort to promote original expression.
  • A SynthID watermark is embedded in all AI-generated tracks for identification, and a system is being introduced to check if tracks are AI-generated.
  • AI-generated music is sparking debate, with platforms like YouTube and Spotify increasingly incorporating AI while facing legal challenges from the industry over copyright issues.
  • Deezer has developed tools to mark and manage AI-generated music to prevent issues like fraudulent streams.