Google and Universal Music in talks to license artist voices for AI music

Just months after a viral AI-generated Drake clone sent record executives into full panic mode, the music industry is trying a new strategy: monetize it before it kills them.

According to the Financial Times, Google is in talks with Universal Music Group to license artists’ voices and melodies for an AI-powered music creation tool. Warner Music is also reportedly discussing a similar deal with the tech giant.

Licensing the voice instead of sending legal threats

Under the proposed framework, copyright holders would be paid for the right to generate music, while individual artists would have the choice to opt in.

The tool would allow fans and creators to legally produce songs using an artist’s voice, sound, or lyrics, rather than relying on unauthorized deepfakes scraped from the open web.

Warner Music CEO Robert Kyncl gave away the industry's changing playbook on an earnings call Tuesday, saying that with the right framework in place, AI could enable fans to pay heroes the ultimate compliment through user-driven content.

Not everyone is convinced. Universal’s roster includes heavyweights like Taylor Swift, Ariana Grande, and Drake, who publicly declared that deepfakes using his vocals were "the last straw." Rock legend Sting told the BBC that human capital faces an inevitable battle against AI, adding that synthetic music doesn't impress him at all.

Others are leaning in. Grimes previously offered a 50-50 royalty split to anyone using her AI-cloned voice, while Paul McCartney revealed in June that AI was used to isolate John Lennon’s voice from an old demo cassette to finalize a new Beatles track.

The scramble for legal guardrails

Universal previously tried locking the front door. Back in April, the label sent letters to streaming platforms demanding they block AI scrapers from harvesting protected lyrics and melodies.

In a Senate hearing last month, Universal general counsel Jeffrey Harleston warned lawmakers that an artist’s voice is the most valuable part of their livelihood and public identity. Meanwhile, the U.S. Copyright Office has stood firm, refusing to grant copyright protection to machine-generated works without human authorship.

Tech platforms, however, are moving ahead regardless. Google released an experimental text-to-music AI tool back in May, and Meta rolled out its own model earlier this month, trained on 20,000 hours of owned and licensed music.

Why it matters

For creators and casual fans, this could turn AI mashups from copyright infringement into an official feature. Instead of having a track scrubbed from TikTok overnight, users might soon pay a fee to borrow a pop star’s vocal cords legally. But don't expect it to be cheap, as premium voices will likely sit behind steep paywalls.

For the music business, this looks like the streaming playbook all over again. Labels spent years fighting Napster and file-sharing before realizing the only way to beat digital piracy was to license it and collect royalties at scale. By partnering with Google, Universal and Warner are setting the industry standard, forcing independent AI music startups to either license or face annihilation in court.

We think labels are taking the only pragmatic route available, but the opt-in clause will make or break the product. If top-tier icons refuse to participate while second-tier artists jump in, the tool could feel like a karaoke gimmick. More dangerously, if licensed AI tracks flood the market, labels risk devaluing the very human brands they spent decades building.