Apple and Google Add Generative Music AI Features
Apple and Google are both rolling out music-focused generative AI features, reflecting a broader push to embed creative intelligence into their platforms. The new capabilities, found in Apple Music and through Google's Gemini, signal a new competitive front in AI. For Apple, this push drives further optimization of its Neural Engine and ML pipelines on Apple Silicon for on-device inference.
- Google's Lyria 3 model, integrated into the Gemini app, generates 30-second music tracks from text, images, or videos, complete with AI-generated lyrics and cover art. This service is available in eight languages, with usage limits depending on the user's subscription tier. - A key feature of Google's approach is the SynthID watermark, an imperceptible audio marker embedded in all AI-generated tracks to ensure transparency and allow for identification. The model is designed to avoid replicating the specific styles or voices of particular artists. - Apple's "Playlist Playground" in Apple Music, a feature within iOS 26.4, uses Apple Intelligence to create 25-track playlists with custom art and descriptions based on user prompts. This reflects Apple's broader strategy of integrating AI as a background tool to augment existing features rather than as a standalone generative function. - Apple’s AI strategy prioritizes on-device processing, leveraging its Foundation Models Framework and the Neural Engine in Apple Silicon to ensure user privacy and reduce reliance on the cloud. The Neural Engine is a specialized AI accelerator integrated into Apple's A-series and M-series chips since 2017, designed for machine learning tasks. - For developers, Apple provides access to its on-device models through the Core ML framework, enabling features like structured outputs and tool-calling within apps without external server dependencies. This approach contrasts with Google's cloud-based processing for Lyria 3, which is too large to run locally. - In the professional space, Apple has integrated AI-powered "Session Players" into Logic Pro, including a virtual Bass Player and Keyboard Player that can automatically accompany a user's music. This builds on their decade-old generative "Drummer" feature. - The generative AI music market is a fragmented and competitive space with numerous startups like Suno and Udio, alongside major tech companies. This has led to legal challenges, with major record labels suing some startups over alleged copyright infringement in their training data. - Google also offers Lyria RealTime, an experimental API for developers that allows for the interactive and continuous generation of instrumental music via a low-latency streaming connection. This enables real-time steering and blending of musical styles.
