Why Warner Music Just Bought an AI Company to Track How Algorithms Use Songs

Warner Music Group has acquired Sureel AI, a startup that builds technology to track how AI models learn from and use artists' music, according to Music Business Worldwide.

The core problem Sureel is solving is simple but crucial: right now, when a company trains an AI model on thousands of songs, there is no clear way to prove which songs were used or how much they influenced what the AI learned. Most arguments about AI and music copyright have focused on whether this kind of training is even legal. Sureel approaches it differently. Instead of asking "was this stolen?" the company asks "what was actually learned from this song, and how much did it matter?" That is a more practical question because it opens a path to payment.

Record labels have been trapped between two bad options: sue AI companies and risk damaging business deals that could make money, or license songs broadly and accept payment terms that might be unfairly low, given that nobody knows what AI-generated music will be worth in a few years. Tools that track what AI models learned from what songs offer a third way—a technical foundation for deals based on actual usage rather than a one-time license payment.

There is a lesson in history here. The music industry fought for years against iTunes and Spotify instead of learning how to profit from them. By the time they cooperated, companies like Spotify already set the rules and the payment rates. The industry has been frustrated with how little artists earn per stream ever since. Owning the tracking technology before everyone agrees on standards is a smarter approach.

From a technical standpoint, this is genuinely hard. If you want to know whether a particular song was in the data used to train an AI, you cannot be completely sure—you can only make educated guesses. Methods exist to measure how much one song shaped an AI's behavior, but they are slow and still improving. Sureel has not shared publicly how it does what it claims to do, so it is hard to know whether WMG bought a finished tool, a promising team, a new technique, or all of these.

Warner did not disclose how much it paid for Sureel.

What this move signals is that the music industry is starting to think differently. Instead of just suing when AI companies use their songs, rights holders are building tools to track and measure that use before legal battles happen. If these tools become reliable enough, AI developers could produce a detailed report: "Here is what was in our training data, here is how much it shaped what we built." That would make it possible to negotiate deals based on facts rather than fighting in court. It would help AI companies know exactly what they can do legally. It would mean artists get paid based on what AI actually learned from their songs, rather than waiting for a lawsuit to settle. That would be healthier for everyone involved.

The big question is whether Sureel's technology can actually work at the scale needed. Tracking this stuff is advancing fast, but it is not yet at a point where anyone can guarantee it will work perfectly. WMG needs to be careful not to oversell what these tools can do, or it may find itself defending promises the science cannot yet keep.