Folk musician Murphy Campbell found AI copies of her songs on Spotify

In January 2026, folk musician Murphy Campbell opened her Spotify profile and saw tracks that she had not uploaded. The songs were familiar — her own recordings — but the vocals sounded strange, as if someone had run the voice through a filter and broke something in the process. After investigating, Campbell came to a disheartening conclusion: someone had downloaded her performances from YouTube, processed the voice using AI voice cloning tools, and uploaded the resulting tracks to streaming platforms under her name.

One of such tracks — "Four Marys" — journalists from The Verge ran through two independent AI content detectors. Both returned the same verdict: the vocals were very likely generated by artificial intelligence. For the artist herself, this was a shock.

According to her, she was confident that there was at least some protection against this. It turned out there wasn't. Voice cloning technology has been available to a wide audience for several years, and the legal system simply hasn't kept up with this pace.

Uploading other people's AI versions of tracks to Spotify is technically straightforward: the platform does not require proof of authorship when publishing through distributors. Campbell's situation is not a one-off case. Over the past two years, musicians around the world have recorded similar incidents: their voices are cloned, their names are used without permission, and the tracks earn streaming royalties that go to scammers.

Independent artists are particularly vulnerable — they lack the legal teams and resources of major labels to track and dispute such violations. The DMCA system, which is traditionally relied upon for copyright protection on the internet, was created for a completely different threat landscape. It assumes that the violator uploads someone else's content as is.

But when AI generates a new file based on someone else's voice — the legal picture becomes blurred. Is a synthetic voice a copyright infringement? Who bears responsibility: the one who trained the model, the one who generated the track, or the platform that accepted it?

There is no clear answer in the legislation of most countries yet. Murphy Campbell's story exposes a double problem: technologies are developing faster than law, and platforms are not yet taking proactive responsibility for content verification. While lawmakers deliberate, ordinary musicians find themselves alone with tools that can create a deepfake of their career in minutes and place it where millions listen.