AI Can Make Unlimited Music. Intelligence is the bigger opportunity.
- Jeff Yasuda
- 3 days ago
- 5 min read
AI can generate a credible song in seconds. Fuzz Collective's Jeff Yasuda considers what that means for musicians, listeners, authenticity, and the future of music.
There is nothing artificial about the photograph above.
It captures Fuzz Collective performing in front of our friends at Brick & Mortar Music Hall in San Francisco. The musicians are responding to one another, the audience is having fun, and everyone in the room is sharing a moment that will never happen in exactly the same way again.
That feeling is one of the things I love most about making music.
It is also why the rapid growth of AI-generated music raises such interesting and uncomfortable questions. If a song created by a machine genuinely moves you, does it matter who, or what, made it? Would you feel differently if you discovered afterward that no human artist existed?
I think about these questions from two perspectives: as a musician with Fuzz Collective and at Feed.fm, where we work at the intersection of music, technology, licensing, and listener experience.
AI can make music. That question has been answered.
Generative AI platforms can now produce credible recordings from text prompts in seconds. The quality is improving quickly, production costs are approaching zero, and the volume of synthetic music is exploding.
In July 2026, Deezer reported receiving approximately 90,000 fully AI-generated tracks each day, with fully AI-generated recordings accounting for more than half of its new uploads on peak days in June. That is not a distant prediction. It is already happening.
For working musicians, the reaction is understandable. Songs that take months or years to write, arrange, rehearse, record, mix, and release are entering the same ecosystem as recordings that can be generated almost instantly and at virtually unlimited scale.
But unlimited supply does not create unlimited attention. It creates a problem of judgment. Listeners and platforms must decide what is authentic, relevant, properly licensed, and genuinely worth hearing.
Does authenticity change how a song feels?
Most music fans do not experience a song as a collection of technical attributes. We connect with a voice, a lyric, a groove, a memory, a place, or a moment in our lives.
Part of that connection can come from knowing there is a person behind the music. We follow artists because their choices, histories, personalities, and imperfections become part of the experience. Discovering a new band can feel like finding something that belongs to you before the rest of the world catches on.
AI complicates that relationship. A synthetic song may still sound good. It may even trigger a genuine emotional reaction. But listeners should be able to know how it was made, whether a real artist's voice or identity was imitated, and whether the people whose work helped create it gave their permission and shared in the value.
The issue is not simply whether AI music is good or bad. It is whether the surrounding system is honest and fair.
The bigger opportunity is music intelligence
The first wave of AI in music has focused on creation. I believe the larger opportunity is intelligence: helping us understand music, listeners, and context more deeply.
Traditional music services rely on signals such as genre, artist, tempo, popularity, and listening history. Those signals are useful, but they only describe part of what makes a song right for a particular person at a particular moment.
A track that works during the final minute of a difficult workout may be wrong during the warmup. Music that helps someone concentrate may not help that same person relax. A familiar song may be far more motivating than a newly generated one when someone needs an emotional lift.
AI can help analyze mood, intensity, lyrical meaning, instrumentation, cultural context, energy progression, and audience suitability. It can also consider what someone is doing, where they are in an experience, and what response the product is trying to create.
That moves us beyond personalization toward adaptation.
Personalization asks, "What music does this person usually like?" Adaptation asks, "What music will best serve this person right now?"
Technology still needs human judgment
Music carries meaning that cannot always be inferred from audio characteristics or metadata. A lyric can be technically permissible and still be wrong for an audience, brand, or moment. A model can identify patterns at enormous scale, but taste, subtlety, culture, and intent remain difficult to automate.
In my company, our curators, data experts, and product teams work together to understand how a track sounds, where it belongs, and how listeners respond. Technology helps us analyze catalogs and behavior. Experienced people provide context and judgment.
The listener matters too. Even perfect metadata cannot tell us everything about what one person needs at a particular time. The best systems will combine algorithmic intelligence with human expertise and real listener response.
The future is not AI replacing human curation. It is AI amplified by human judgment.
Artists must remain part of the value
AI is forcing the music industry to confront questions about training data, artist consent, vocal identity, attribution, and compensation. These are no longer abstract policy debates. They are becoming product requirements.
The sustainable path is not to treat recorded music as an unlimited pool of raw material. It is to create systems in which artists, songwriters, rights holders, technology companies, and listeners can all participate in the value being created.
That means transparency when music is generated by AI, consent when an artist's work or identity is used, compensation when value is created, and meaningful protection against fraud and impersonation.
Technology can help music reach more people and become more responsive inside fitness, wellness, gaming, healthcare, and other digital experiences. But innovation should not require disconnecting music from the people who made it possible.
Music is more than sound
For musicians, AI presents real risks. It can flood platforms with synthetic content, imitate artists, complicate attribution, and make it even harder for human work to be discovered.
But AI can also help the right music reach the right listener at the right moment. Used responsibly, it can deepen discovery and create new experiences without severing music from the people who make it.
I do not believe the future of music will be entirely synthetic or entirely human. The more interesting future combines technology with taste, context, permission, and human judgment.
As a musician, that last part matters most to me. Music is more than sound. It is expression, identity, community, and the connection created between people when a song finds its audience.
What do you think? If an AI-generated song genuinely moved you, would knowing how it was made change the experience?
If you are interested in hearing more, I will be speaking about these topics at the Dartmouth Entrepreneurs Forum in San Francisco on September 11, 2026, hosted by the Magnuson Center for Entrepreneurship at Dartmouth.
Dartmouth Entrepreneurs Forum link:
This essay was adapted from a longer article originally published on LinkedIn:
Comments