The Legal Divide in AI Music: ElevenLabs vs. Suno

The world of AI-generated music is moving fast and now, it’s getting complicated. On one side, companies like Suno train their models on vast oceans of songs, many of which might not be officially licensed. On the other, new players like ElevenLabs are trying to build AI music generators that stay inside the legal lines.

Earlier this year, ElevenLabs announced a new model that can generate music cleared for commercial use (TechCrunch, 2025). This marks a big step for the company, which was previously known for its text-to-speech and voice synthesis tools. With this move, ElevenLabs is stepping into a crowded space but with a more cautious, rights-friendly approach.

To build its model, ElevenLabs partnered with Merlin Network and Kobalt Music Group, two major publishing platforms that represent artists like Adele, Beck, and Phoebe Bridgers. Musicians who work with these companies can opt in to allow their songs to be used for AI training meaning they share in the revenue that comes from it. It’s a fairer, more transparent system than what some other AI platforms use, and it’s designed to protect artists’ rights.

By contrast, tools like Suno have faced lawsuits from the Recording Industry Association of America (RIAA) over allegations that their models were trained on copyrighted music without permission. That gives Suno a more controversial reputation, though it also means its models may have access to a much broader range of sounds, styles, and genres.

For AI musicians, this creates a tricky choice. A “legal-first” tool like ElevenLabs offers peace of mind: you can release music commercially without worrying about lawsuits or takedown notices. The trade-off, however, is variety. Because ElevenLabs only trains on licensed or volunteered music, its creative range may be smaller. Meanwhile, Suno’s enormous (and less restricted) dataset can produce more diverse results, but comes with legal gray areas and ethical concerns.

In simple terms, it’s the classic balance between freedom and fairness. The safer the data, the narrower the palette. The wilder the training, the greater the risk. Every AI artist will have to decide which side of that line they want to stand on.

In the end, ElevenLabs’ entry shows that the AI music industry is growing up. The future of AI-generated sound may not just be about who can make the best track: but who can make it legally...