Google Gemini Gains AI Music Generation via Lyria 3
Google is integrating advanced music-generation capabilities into the Gemini app, leveraging its latest DeepMind Lyria 3 model. This beta rollout allows users worldwide to create custom tracks and lyrics simply by providing descriptive prompts.
How the Lyria 3 Integration Works
The feature functions through text-based prompts where users specify the type of song they wish to produce. For example, requesting a “comical R&B slow jam about a sock finding its match” results in a 30-second audio track accompanied by cover art generated by Nano Banana. Beyond text, the tool can analyze uploaded photos or videos to compose music that matches the specific mood of the media.
Lyria 3 represents a significant step forward from previous iterations, offering increased complexity and realism in musical output. Users gain granular control over the creative process, with the ability to adjust:
- Musical style
- Vocal characteristics
- Tempo and rhythm
Global Expansion and Creator Tools
Alongside the Gemini update, Google is expanding its “Dream Track” feature to YouTube creators on a global scale. Previously restricted to U.S. users, this tool enables creators to leverage AI for their content production. While the model allows users to reference specific artists in their prompts, Google clarifies that this is intended for stylistic inspiration rather than direct imitation.
“Music generation with Lyria 3 is designed for original expression, not for mimicking existing artists,” the company stated in a blog post. To ensure compliance, Google has implemented filters that cross-reference generated outputs against existing copyrighted content.
Safety and AI Identification
To address concerns regarding the authenticity of digital audio, all tracks produced by Lyria 3 will feature a SynthID watermark. This technology allows for the identification of AI-generated content. Furthermore, Gemini is gaining the ability to analyze uploaded tracks to determine if they were created using AI, providing users with a verification mechanism.
This functionality is now available to Gemini users aged 18 and older, with language support spanning English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese. The launch arrives amid a broader industry debate, as platforms and labels navigate the complexities of AI training data, copyright litigation, and the effort to curb fraudulent streaming practices.