AI Lets Anyone Sing Like a Star Using Their Own Voice
Creatrip Team
6 days ago
AI music services now let users train a model on a short recording (about 4 minutes) of their own voice so the AI can sing new songs in that voice. US service Suno added a paid voice feature that creates full songs and music videos from prompts and user-recorded vocals. AI-produced singers (e.g., 'Jania Mone' and 'Breaking Rust') have already topped Billboard charts, and Google introduced its own music generator, Lyria3 (via Gemini). The technology speeds song creation, and streaming platforms report a surge of AI tracks—Dizer said AI tracks made up 34% of new releases at one point—prompting some services like Bandcamp to ban AI uploads. Copyright fights continue: Warner Music settled with Suno to share revenue for songs using its artists' voices, while in Korea the Music Copyright Association (음저협) has paused registering AI-created works until clearer laws and detection tools exist. Industry groups formed a 'K Music Rights Coexistence Committee' (K음악권리단체 상생위원회) to seek solutions, but policy and enforcement remain unresolved.
