Google launches Lyria 3 Pro to automate full song production

New AI model allows users to generate three-minute tracks with distinct structural elements

Google launches Lyria 3 Pro to automate full song production

The landscape of the music industry has shifted dramatically with the release of Lyria 3 Pro on Wednesday. Developed by Google DeepMind, this advanced AI model moves beyond simple loops to generate high-fidelity, three-minute songs. Unlike its predecessors, Lyria 3 Pro allows creators to define specific musical structures—such as intros, verses, choruses, and bridges—through natural language prompts or even images.

While Google emphasises that the tool is designed to assist human producers, the ability to automate entire compositions has intensified the debate over the future of human musicianship.

The scale of AI integration is staggering; Deezer recently reported that nearly 40% of all daily uploads are now fully AI-generated, equating to roughly 60,000 tracks every 24 hours.

A joint study by Deezer and Ipsos found that 97% of listeners were unable to distinguish these synthetic tracks from human-made music. This "invisibility" has led to a surge in fraudulent activity.

On 23 March 2026, Michael Smith, a 54-year-old from North Carolina, pleaded guilty to defrauding streaming platforms of over $8 million.

Smith utilised AI to generate hundreds of thousands of songs and deployed a network of 10,000 bots to stream them billions of times, effectively stealing royalties meant for real artists.

In response, the industry is racing to build defensive technology. Spotify has launched a pilot for "Artist Profile Protection" to prevent AI deepfakes from being misattributed to popular stars.

Meanwhile, Deezer has begun licensing its proprietary AI-detection technology—which boasts 99.8% accuracy—to other platforms and rights organisations like Sacem.

Despite these safeguards, the PRS for Music notes that 74% of its members remain deeply concerned about AI competition.

The consensus among experts is that while AI can replicate the mechanics of a song, it still lacks the lived experience and emotional depth that define human artistry—though the economic pressure on traditional composers has never been higher.