Google’s Gemini app gained a new capability on Wednesday, February 18, 2026: the ability to generate original music from text prompts, images, and videos. The feature, powered by Google DeepMind’s Lyria 3 AI model, allows users to create 30-second music tracks with custom lyrics, marking a significant expansion of Gemini’s creative toolkit.
Unlike specialized AI music platforms like Suno or Udio, which operate as standalone services, Lyria 3 is integrated directly into the Gemini chat interface. This allows users to seamlessly generate music within their existing workflow, according to a blog post from Google DeepMind.
Lyria 3 represents a substantial technological leap forward, capable of autonomously generating both lyrics and vocals based solely on user intent or mood. Users can input a text prompt – for example, “a comical R&B slow-jam about a lost sock” – or upload a visual, such as a photograph of a sunset, and the AI will compose a corresponding soundtrack. The system also allows for granular control over musical parameters, including genre, tempo (BPM), and sonic density through what Google calls “weighted prompts.”
In response to concerns about the proliferation of AI-generated content, Google has implemented SynthID, a digital watermark embedded in every track created by Gemini. This inaudible marker verifies the artificial origin of the music, addressing anxieties within the music industry regarding the potential for platforms to become saturated with unidentifiable synthetic audio.
Google asserts that Lyria 3 is designed to facilitate “original expression” and prevent the replication of existing artists’ voices. The model incorporates filters to prevent the cloning of voices, and when prompted to create music “in the style of” a particular artist, it will draw inspiration from their overall style without directly copying their vocal or melodic signatures.
Despite these safeguards, the legal status of AI-generated works remains ambiguous as of February 2026. Current legal directives stipulate that raw files generated by Gemini are not eligible for copyright protection. To establish ownership, artists must demonstrate “significant human authorship” by substantially editing the AI-generated stems using digital audio workstations (DAWs). This suggests a future where AI serves as a prototyping tool for musicians, accelerating the demo creation process but still requiring human intervention to achieve commercial viability.
A key application for Lyria 3 is its integration with YouTube Shorts. Google aims to empower creators with the ability to generate unique, royalty-free soundtracks for their short-form videos, bypassing the often-complex licensing requirements associated with social media platforms. The feature is currently available globally to users aged 18 and over, effectively transforming millions of smartphones into portable music studios.