Google’s Gemini Now Creates AI Music with Lyria 3

Google has integrated artificial intelligence-powered music creation into its Gemini virtual assistant. The new capability, powered by the company’s next-generation music model, Lyria 3, allows users to generate 30-second music tracks from text or image prompts.

Lyria 3: AI-Powered Music Generation

Lyria 3 transforms text or image prompts into complete 30-second tracks, including instrumentals, vocals, and lyrics . Users can describe the desired song, specifying genre, mood, or even referencing personal memories, and Gemini handles the technical complexities. The model is designed to explore various languages and genres and can even create music based on uploaded images. The resulting audio quality aims to meet professional standards .

SynthID: Watermarking AI-Generated Content

To address concerns about the origin of AI-generated content, Google is embedding an invisible digital watermark called SynthID into every track created through Gemini. This allows for the identification of content generated by Google’s AI. Google is expanding its verification tools, enabling users to upload audio files and determine if they were created by Google AI .

Copyright and Training Data

Google has stated that Lyria 3 was trained using music that YouTube and Google have the legal rights to use, adhering to their terms of service, partner agreements, and applicable regulations . The company acknowledges the importance of copyright and has implemented filters to check outputs against existing content. However, if a user specifies an artist in a prompt, Gemini will interpret this as broad creative inspiration, creating a track with a similar style or mood .

Industry Response and AI Music Concerns

The launch of Lyria 3 occurs during a broader debate regarding AI’s impact on the music industry. Streaming platforms are taking steps to address AI-generated music. Apple Music has stopped monetizing 2 billion fraudulent streams, representing $17 million in royalties . Deezer has similarly demonetized 85% of AI-generated songs on its platform, utilizing a specialized detection tool . Bandcamp has banned AI-generated music altogether .

Studies indicate that 97% of respondents cannot distinguish between human-created and AI-generated music. Estimates suggest that music sector workers could lose up to 25% of their income in the next four years due to AI advancements . Concerns have been raised by prominent musicians, including Paul McCartney, Kate Bush, and Elton John, who have urged governments to protect creative works from AI’s impact.

Availability

Lyria 3 music generation in Gemini is currently available globally in English, Japanese, Korean, Hindi, Spanish, Portuguese, German, and French for users over the age of 18 . Access to the feature is automatic for users with access to the Gemini app. Administrators can control access through the Generative AI settings in the Workspace Admin console .