Google Brings AI Music Generation to the Masses with Lyria 3 in Gemini

Henry Wadsworth Longfellow once called music “the universal language of mankind.” But what happens when that language is being generated not by human hands on instruments, but by a probabilistic machine learning model analyzing patterns in data? Google is about to find out on a massive scale.

This week, the company announced that its latest Lyria 3 AI music model is rolling out directly in the Gemini app, bringing AI-generated music to one of the world’s most popular AI platforms. It’s a move that could fundamentally change how millions of people interact with creative tools—and raise fresh questions about the nature of art in the age of algorithms.

“The universal language of mankind” — Henry Wadsworth Longfellow, on music

From Developer Tools to Consumer Apps

Google DeepMind has been quietly developing Lyria for years, offering limited access through developer-oriented products like Vertex AI. But Lyria 3 represents a significant leap—not just in capability, but in accessibility. Now, anyone with a Gemini account can generate music with a simple prompt.

The process is deliberately simple. Users select the new “Create music” option in the Gemini app or web interface, describe what they want, and optionally upload an image to help set the right vibe. Within seconds, they receive a 30-second track complete with automatically generated lyrics.

Unlike previous versions that required users to provide their own lyrics, Lyria 3 can handle vague prompts and create suitable lyrics on its own. The system also generates an album cover-style image for each track using Google’s Nano Banana model. For those seeking inspiration, Gemini comes pre-loaded with AI tracks that users can remix to their heart’s content.

The Sound of AI Innovation

Genre versatility appears to be one of Lyria 3’s standout features. Google has showcased examples ranging from nostalgic afrobeat tracks about childhood memories to gritty 1970s Motown soul parodies, wistful drum-and-bass pop, and even authentic a cappella sea shanties with robust male choirs and wooden deck foot-stomps.

YouTube integration extends the model’s reach beyond Gemini. The Lyria 3 tools are also coming to Google’s Dream Track toolkit for YouTube Shorts, where they’ll pair with the Veo AI video generation options. This creates a comprehensive AI content creation ecosystem within Google’s platforms.

Watermarking and attribution represent Google’s attempt to address the inevitable authenticity concerns. Every track generated with Lyria 3 will have an audio version of Google’s SynthID embedded within it. Users can upload any audio file to Gemini to check if it was created with Google’s AI—similar to the existing system for images and videos.

“We’re about to see a lot more AI music on the Internet.” — Industry observers on Google’s Gemini integration

Copyright, Creativity, and the Road Ahead

Google says it has sought to build a music AI that respects copyright and partner agreements. If users name a specific artist in their prompt, Gemini won’t attempt to copy that artist’s sound. Instead, it’s trained to treat such references as “broad creative inspiration.”

But the company acknowledges this process isn’t foolproof. Some generated tracks might still imitate original artists too closely. In those cases, Google invites users to report such content—a reactive rather than preventive approach that may not satisfy critics of AI-generated content.

The launch raises fundamental questions about the future of music creation. Streaming services have already seen an influx of AI-generated artists, some gathering thousands of listeners who may not realize they’re listening to algorithmic output. By bringing these capabilities to Gemini—one of the most accessible AI platforms—Google is poised to dramatically accelerate this trend.

Lyria 3 is rolling out now in the Gemini web interface, with mobile app availability expected within days. The feature supports English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese, with more languages planned. While all users get some access, those with AI Pro and AI Ultra subscriptions will have higher usage limits.

This article was reported by the ArtificialDaily editorial team. For more information, visit Ars Technica.