Google Expands Gemini With AI Music Generation

Google has enabled its Gemini AI assistant to create music through its new Lyria 3 model.

Google has introduced music generation capabilities to its generative AI, Gemini AI assistant through the Lyria 3 model, which the company describes as its most advanced music generation system to date. The beta version allows users to generate 30-second tracks using text, image, or video prompts, producing compositions complete with instrumentals, vocals, and lyrics.

Lyria 3 is designed as a high-fidelity AI music generator capable of transforming a specific feeling, image, or concept into a custom soundtrack. Google notes that the system can blend different genres, moods, and eras, creating a wide variety of musical outputs. Access to Lyria 3 is currently restricted to users over 18 via the Gemini app.

To address authenticity and content tracking, all music generated through Gemini includes SynthID, Google’s imperceptible watermark for AI-generated audio. Users can also upload audio files to Gemini to verify if content was produced with Google AI, with the assistant applying SynthID detection and internal reasoning to confirm results.

This development comes amid growing scrutiny of AI-generated music on streaming platforms. Apple Music, for instance, recently demonetised two billion AI-generated streams, representing approximately $17 million in royalties, according to Oliver Schusser, Apple Music’s Vice President. Deezer has similarly removed 85% of AI-generated tracks from its platform using its proprietary AI-detection tool, which identifies content from models such as Suno and Udio, noting that it now processes 60,000 AI-generated tracks daily.

Google has not disclosed the specific training data or models used for Lyria 3. However, the company emphasizes the system’s ability to mix multiple musical styles and eras while maintaining high fidelity. The integration of music generation into Gemini represents a strategic expansion of Google’s artificial intelligence capabilities, linking creative applications with its existing AI verification technologies.

Google’s move positions the company alongside other technology firms exploring AI-driven music creation, even as the industry navigates the challenges of monetisation, copyright, and content verification.

The Gemini app’s music generation feature is now available to select users, allowing them to experiment with short-form AI-composed tracks, while Google continues to refine its AI tools for both creative production and content authentication.