Google and Apple are ramping up their AI ambitions with new music-focused features embedded in their flagship consumer apps.

Alphabet-owned Google has begun rolling out a music-generation tool inside its Gemini assistant, powered by DeepMind’s latest Lyria 3 model.

The feature allows users aged 18 and over to generate 30-second tracks from text prompts, as well as uploaded photos or videos. Users can request custom lyrics or instrumental-only audio, and specify genre, mood, tempo and vocal style.

The tool is launching first on Gemini’s desktop app, with mobile availability to follow in the coming days. Supported languages include English, German, Spanish, French, Hindi, Japanese, Korean and Portuguese.

Google says tracks created in Gemini will include AI-generated cover art via its Nano Banana image model, making them easier to share. The company has also embedded its SynthID watermarking technology to identify AI-generated audio.

Free users will be able to create up to 10 tracks per day, while paid subscribers can generate between 20 and 100, depending on their plan. Google says it has implemented safeguards to prevent direct imitation of specific artists, with prompts referencing musicians treated as “broad creative inspiration” rather than instructions to replicate a voice or style.

Meanwhile, Apple is introducing its own AI-driven music feature in the form of “Playlist Playground” within Apple Music.

The tool, included in the iOS 26.4 beta, enables users to generate playlists from text prompts. Each playlist includes 25 songs, along with AI-generated cover art and descriptions. A wider rollout is expected later this year.

The moves place both tech giants in more direct competition with Spotify, YouTube Music and other streaming platforms that have already introduced AI playlist tools.

The features arrive amid ongoing scrutiny from the music industry over AI training practices and copyright protections.

In 2024, Universal, Sony and Warner sued AI startups Suno and Udio over alleged use of copyrighted recordings in training data.

Since then, licensing deals have started to emerge, with Warner settling with Suno last year under an opt-in model allowing artists to license their voices and likenesses.

Sony has also developed a tool aimed at tracing AI-generated tracks back to original works, potentially enabling royalty allocation based on influence.