Google Adds AI Music Creation to Gemini with Lyria 3 Model

Google has introduced a new AI-driven music creation feature within its Gemini application, marking a significant step in generative artificial intelligence for audio. Announced on Wednesday, the capability is powered by DeepMind’s Lyria 3 model and is now available to users worldwide aged 18 and older.

The music generation tool is currently in beta and enables users to produce 30-second musical tracks by describing the desired style or mood. Users can input prompts such as “a comical R&B slow jam about a sock finding its match,” and the AI will generate an original song complete with lyrics and visually generated cover art using a tool called Nano Banana.

Beyond text-based prompts, the Gemini app now supports media uploads. Users can submit photos or videos, which are analyzed by the AI to assess the emotional tone. The application then creates a song intended to match the mood depicted in the uploaded file, broadening creative possibilities for both casual users and content creators.

The Lyria 3 model represents the latest progress in Google’s AI audio research, delivering more realistic and complex musical outcomes than previous iterations. Individuals can adjust aspects such as musical style, vocals, and tempo, offering greater control over the generated output. While users may reference specific artists in their prompts, Google has clarified that the system is designed to foster original expression rather than imitation. The AI uses artist names strictly as stylistic inspiration, with safeguards in place to prevent copying or replication of real artists’ voices or sound.

The release follows a broader trend within the music and technology industries as platforms and companies balance creative innovation with intellectual property concerns. To address this, all audio created with the Lyria 3 model includes a SynthID watermark, which verifies that the track was generated by AI. The Gemini app is also being upgraded with features allowing users to check whether audio uploads are AI-created, utilizing the same watermarking technology for identification.

Google has expanded the Lyria 3 model’s availability beyond Gemini by rolling it out globally to YouTube creators via the Dream Track feature. Previously restricted to YouTube creators in the United States, this update provides a wider group of content creators with the ability to craft original music tracks for their videos using AI technology.

The rise of AI-generated music comes as companies like YouTube and Spotify enter licensing discussions and partnerships with major music labels to monetize synthetic content. At the same time, AI developers face ongoing legal challenges from the music industry over training data and copyright usage. Other platforms have adopted their own measures; for example, Deezer has implemented tools to mark AI-generated songs to deter fraudulent streaming. Google’s incorporation of watermarking and detection tools reflects growing attention to both creativity and the responsible management of synthetic media in the digital ecosystem.

dataconomy.com