Lyria 3 Pro Extends AI Tracks to 3 Minutes
Google announced on March 25 its Lyria 3 Pro model, a major upgrade to its music generation AI that extends the maximum track length from 30 seconds to three minutes. This sixfold increase allows the technology to move beyond creating short clips to composing full songs. The new model also enables users to specify musical structures such as intros, verses, and choruses, granting granular control over the final arrangement. This enhancement positions Lyria to compete directly with specialized AI music platforms like Suno and Udio, which have gained traction by offering similar long-form generation capabilities.
Google Embeds AI Music Across Enterprise and Consumer Platforms
In a clear signal of its commercialization strategy, Google is integrating Lyria 3 Pro across its entire product suite. For enterprise customers, the model is now available in public preview on Vertex AI, allowing businesses to generate high-fidelity audio at scale for applications like custom game soundtracks. Developers can access the model through the Gemini API and AI Studio. For consumers and creators, the technology is being rolled into the Gemini app for paid subscribers, the video creation tool Google Vids, and the recently acquired AI music platform ProducerAI. This multi-pronged deployment leverages Google's existing infrastructure and user base to drive adoption and create new revenue streams from its AI research.
SynthID Watermark Addresses Copyright Concerns
Anticipating legal and ethical challenges surrounding AI-generated content, Google is embedding its SynthID watermark into all audio produced by Lyria 3 Pro. This inaudible marker identifies the content as AI-generated, providing a mechanism for transparency and attribution. Google stated the model is trained on a combination of licensed partner data and publicly available content from its platforms. While the company claims Lyria does not directly mimic specific artists, it acknowledges that prompts naming an artist will be used as "broad inspiration." This proactive watermarking and carefully worded policy aim to mitigate copyright infringement risks as generative AI tools become more powerful and widely adopted.
