Google Integrates Lyria 3 AI Music Generation into Gemini App

Google's Lyria 3 Integrates into Gemini App

Google has integrated its Lyria 3 generative music model into the Gemini app, enabling users to create 30-second audio tracks using text and visual prompts. This new feature allows for extensive customization of genre, mood, style, vocals, and tempo, and can generate lyrics based on user input. Crucially, all AI-generated audio includes an embedded SynthID watermark for identification and transparency.

Lyria 3 empowers Gemini users to craft personalized 30-second audio tracks, complete with customizable elements and AI-generated lyrics, all while ensuring transparency through embedded SynthID watermarks.

Unleashing Creative Soundscapes with Lyria 3

The Lyria 3 model, developed by Google DeepMind, is specifically designed to generate music and lyrics based on user specifications. Within the Gemini interface, users can initiate music creation by selecting a dedicated "Create Music" button or by providing a descriptive prompt. Input methods are versatile, including textual prompts, uploading photos, or offering other forms of creative inspiration.

Users are granted extensive control over various facets of the generated music, allowing for truly personalized creations:

• Genre: Options span a wide range, from pre-defined categories like 90s rap, Latin pop, R&B romance, and Afropop, to custom descriptions provided by the user.
• Mood and Style: Specific instructions can be given to dictate the overall feel and musical style of the track.
• Vocals and Tempo: Users can decide on the inclusion of vocals and specify the desired speed of the track.
• Lyrics: The model boasts the capability to generate lyrics based on user prompts, even incorporating references from uploaded images or personal memories. For instance, a user could prompt: "Use these photos to create a track about my dog Duncan on a hike in the woods."

Output, Objectives, and Ethical Safeguards

Upon generation, Lyria 3 produces a 30-second audio track, each accompanied by custom cover art. Users have the flexibility to download the generated audio clip or share it easily via a provided link.

Google has stated its core objective for these AI-generated audio tracks: to provide a "fun, custom soundtrack to daily life" and to serve as a powerful tool for original expression. The company has also implemented important safeguards, indicating that the model incorporates filters designed to avoid mimicking existing artists and to detect existing content, thereby aiming to prevent the generation of content similar to copyrighted musical works.

Google's Lyria 3 aims to offer a "fun, custom soundtrack to daily life" and support "original expression," incorporating filters to prevent mimicry of existing artists and content.

Availability and AI Transparency

Lyria 3 is currently rolling out in the Gemini app for users aged 18 and above. The feature is initially accessible on desktop, with plans for mobile app availability in the near future. It supports multiple languages, including English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese, with further language expansion anticipated.

The model's capabilities extend beyond Gemini; it is also accessible through YouTube's Dream Track, currently available in the US and slated for rollout to other countries. This integration allows users to create custom soundtracks specifically for YouTube Shorts or uploaded videos.

All audio tracks generated by Lyria 3 include embedded SynthID, Google's proprietary watermark for identifying AI-generated content. The Gemini app itself will also receive additional tools for identifying AI-generated content, including audio verification. Users will have the ability to upload an audio clip to Gemini to verify if it was generated by the chatbot.

Subscribers to Google AI Plus, Pro, and Ultra tiers will benefit from higher usage limits for the music generation feature, though specific limits for free users were not detailed.

Broader Industry Landscape

The landscape of AI song generation is evolving rapidly, with other platforms like Suno and Udio having previously established their presence. These platforms have notably engaged in licensing agreements with music labels following various copyright challenges. Separately, YouTube further enhanced its offerings by introducing AI Playlist generation for its Premium subscribers earlier in the month.