AI Music's Laundering Problem: What the New Suno Lawsuit Means for Anyone Making Music with AI

What the lawsuit actually argues

Sony and Universal filed their 45-page complaint in a US federal court in Massachusetts, and their argument is direct: Suno's newly licensed v6 model is still built on stolen material.

The theory is that the problem lies not in Suno directly using those recordings in v6, but in the way it built newer models on the outputs of earlier ones that were already trained on unlicensed music. The labels call this approach "model laundering." In their view, taking the results of an infringing model and using those results to train a successor does not erase the original violation. Instead, the protected content is carried forward, embedded in the intermediate outputs and then woven into the new model.

Suno's response

The new front in the litigation focuses on v6, which Suno launched in early September through partnerships with Warner Music Group, BMG and Believe. Suno's Chief Product Officer told Music Business Worldwide that v6 was trained entirely from scratch on different data and without Universal or Sony recordings.

Suno calls the lawsuit "fundamentally flawed," saying its AI is about creating unique music, not copying anyone.

What's at stake

The labels are seeking statutory damages and an injunction. If a court found willful infringement and awarded the maximum amount per work, damages could theoretically exceed $9 billion.

The dispute continues an ongoing legal fight that began in 2024, when Universal, Sony and Warner all sued Suno over its original training practices. Warner later settled and struck a licensing agreement with Suno, leaving Universal and Sony as the remaining plaintiffs.

Why this matters if you use AI music tools

The case raises a question that goes well beyond Suno. If a court accepts the labels' laundering argument, it would mean that any AI model trained even partly on the outputs of an earlier, allegedly infringing model carries legal risk forward, regardless of what licensing deals were struck afterward.

That could affect how you think about any AI music tool you use to generate, arrange or produce your work: not just whether the company behind it has a licensing agreement in place today, but what its earlier models were trained on before those agreements existed.

It also matters in a practical sense for anyone releasing music made with or assisted by AI tools. As legal clarity around these platforms remains unsettled, the provenance of an AI model's training data is becoming a legitimate question to ask before you build a workflow around it.