Extended Track Length and Creative Control

Google unveiled Lyria 3 Pro, a music‑generation model that expands the capabilities of its predecessor, Lyria 3. While Lyria 3 could produce 30‑second clips, Lyria 3 Pro allows creators to generate full tracks lasting up to three minutes. The model also improves understanding of song structure, enabling users to include specific elements—intros, verses, choruses, bridges—directly in prompts for more precise results.

Integration Across Google Products

The Pro model is being integrated into several Google services. Paid subscribers to the Gemini app will receive access, and the model is also being added to Google Vids, the video‑editing platform, as well as to ProducerAI, a generative‑AI music‑production tool Google recently acquired. Enterprise‑focused tools such as Vertex AI (in public preview), the Gemini API and AI Studio will also support Lyria 3 Pro, extending the technology to business users.

Training Data and Artist Inspiration

Google explained that Lyria 3 Pro was trained using data from its partners and permissible content from YouTube and Google. The company emphasized that the model does not copy any specific artist, though it can draw “broad inspiration” when a user references an artist in a prompt. This approach aims to balance creative flexibility with respect for original creators.

AI Identification and Industry Context

All tracks generated with Lyria 3 or Lyria 3 Pro are automatically marked with a SynthID, a label that signals AI involvement. This move aligns with recent industry efforts to provide transparency around AI‑created music. In parallel, platforms such as Spotify and Deezer have introduced tools that let artists verify and identify AI‑generated songs, helping to prevent misattribution.

Used: News Factory APP - news discovery and automation - ChatGPT for Business