Google has integrated DeepMind's latest music model, Lyria 3, into its Gemini AI, enabling users to generate up to 30-second music clips with vocals in any genre. The music is created through text prompts describing how it should sound, with options for 16 preset genres or inspiration from uploaded images. All tracks generated by Gemini are marked with Synth ID, an invisible watermark embedded in the music to identify AI origin.
Science & technology1 min
Google's Gemini AI Now Generates 30-Second Music Clips with Vocals Using Lyria 3 Model
Key Points
- Google's Gemini AI can now generate 30-second music clips with vocals using DeepMind's Lyria 3 model.
- Users create music through text prompts, with options for 16 preset genres or image-based inspiration.
- All generated tracks include Synth ID, an invisible watermark to identify AI-created content.
Transparency
How we verified this article
LowBased on 2 sources
2 sources4 Involved