My Music My Choice tool poisons songs to prevent AI voice cloning and deepfake songs

This invisible technique poisons songs so AI can't clone them

My Music My Choice alters songs so they sound normal to fans but become unusable nonsense when fed into generative AI systems. Last year, AI clones of Bad Bunny and Drake flooded streaming platforms. Listeners couldn't tell the real tracks from synthetic soundalikes. The music industry has been scrambling for answers ever since. Researchers at Binghamton University and the startup Cauth AI think they've found one. It's called My Music My Choice, or MMMC, and it works differently from most copyright tools. Instead of catching fakes after they appear, this method lets artists poison their recordings before release. The audio reaches human ears just fine. But voice cloning models hear nothing but garbage. Here's how the poisoning actually works The system targets a song's waveform. My Music My Choice adds microscopic alterations so subtle that you'll never notice them. Play the track on Spotify and it sounds exactly like the master recording. Recommended Videos But feed that file into cloning software and everything breaks. The shifts confuse the algorithm, making the protected vocals read as a completely different performance. When the tool tries to replicate the voice, it only produces distorted static. The goal is to minimize the impact on human listeners while maximizing disruption for the machines. Artists could apply this protection during production and release with confidence that cloning software won't work. Why last year's wave made this urgent Bad Bunny drops a new track and within hours the internet fills with studio-quality versions sung by anyone. Generative AI made that scenario real in 2025. Fans couldn't tell what was authentic anymore. Beyond the copyright chaos, artists watched their identities get borrowed without permission. People are using voice cloning for fun but also for nefarious purposes, Ciftci said, grabbing someone's voice and making them sing things they never would. The emotional toll and lost revenue piled up fast. Musicians needed a way to shut it down before it starts. MMMC finally gives them that. What's next for artists and the tool The team tested MMMC on 150 tracks across multiple genres and plans to scale up. They also want to compare it with similar methods, though they admit there aren't many out there yet. For musicians watching this space, the message is clear. Protection is coming before the clone, not after. Watch for wider testing as the team scales up.

Deepfake Songs Are Exploding. This Tool Shuts Them Down. | Newswise

Newswise -- Artificial intelligence models can now clone a voice with just a few seconds of audio, fueling a surge of deepfake songs online and creating a growing crisis for musicians who don't want their voices hijacked. Beyond the obvious intellectual property rights issue, this can lead to lost revenue and take an emotional toll on artists who put their heart and soul into their songs. But researchers have a solution. In collaboration with the startup company Cauth AI, faculty and students at Binghamton University, State University of New York have developed My Music My Choice (MMMC), a digital safeguard that empowers artists by protecting their songs from generative AI cloning. Consider this scenario: Bad Bunny has just released a new song, but suddenly, the internet is flooded with countless studio-quality versions sung by famous/infamous people around the world, thanks to generative AI. With everyone able to produce their own high-quality version of the song, even the most diehard fans of Bad Bunny would be hard-pressed to tell the real track from a synthetic imitation. Umur Aybars Ciftci, a research assistant professor in the First-Year Research Immersion Program at Binghamton University, and his collaborator, Ilke Demir, CEO and founder of Cauth AI, want to stop that from happening to today's artists. "Even though this AI technology has been developed for fun and entertainment, a lot of people are using it for nefarious purposes," said Ciftci. "You can easily take someone's voice and make them sing something that they normally don't sing, or steal someone's songs and make it look like it is your song to begin with." My Music My Choice works by adding small, imperceptible changes to a song's waveform. When you play the song back, the vocal will sound exactly the same to your ears. But when an AI model tries to replicate the song, it will only produce distorted noise. From the AI model's perspective, the slight shifts made by My Music My Choice make the protected audio sound like a completely different vocal track - and the AI model struggles to replicate it. "Collaborating with disruptive startups like Cauth AI provides us with a unique vantage point into the front-line challenges of the industry, essentially bridging the gap between lab-scale concepts and industrial-scale impact. Our goal is to build a model that figures out exactly which tiny modifications to introduce so that people hear no difference at all, while AI voice-cloning systems are thrown off," said Ciftci. "In other words, we're trying to minimize the impact on human listeners while maximizing disruption for the machines." If you're a musician with a new track, Ciftci said, this is something you could apply to a song before releasing it to protect it from AI voice cloning. The researchers tested the tool on 150 music tracks across multiple genres, and they will continue testing this system on larger data samples. They also want to compare My Music My Choice with similar methods, though Ciftci said there aren't many out there. Binghamton students Gerald Pena Vargas, Alicia Unterreiner and David Ponce contributed to this research. The paper, "My Music My Choice: Adversarial Protection Against Vocal Cloning in Songs," was presented at the 39th Conference on Neural Information Processing Systems (NeurIPS 2025) Workshop: AI for Music.