Google's Lyria 3: AI Music's New Maestro

Google DeepMind has introduced Lyria 3 and Lyria 3 Pro, its latest multimodal AI music generation models. These advanced tools are designed to create higher-fidelity music with improved structure, aiming to capture a share of the viral AI song market. Lyria 3 integrates with Gemini, offering features like text-to-music generation and AI-generated cover art, with Pro versions providing more advanced song structure capabilities.

The models support multiple languages and can generate audio with emotional inflection, offering flexibility for creators. Prompting guides provide users with a framework for detailed control over genre, mood, and instrumentation. However, the technology is also embroiled in controversy, with a lawsuit filed in March 2026 against Google for allegedly training Lyria 3 on millions of copyrighted music clips without compensation.

Legal experts highlight that AI-generated works are only copyrightable with significant human creative input, not solely from prompts. Google asserts its training data usage is lawful. The company embeds its AI audio with SynthID watermarking and aims to avoid mimicking specific artists, though questions about originality and fair use persist, alongside ongoing legal scrutiny and evolving licensing agreements.