Suno Filed a 44-Page Pushback Against Universal and Sony. The AI Music Copyright War Is Getting Serious.

By Hector Herrera | October 28, 2026 | Creative

Suno filed a 44-page legal answer on October 20 to the second copyright lawsuit brought by Universal Music Group and Sony Music, pushing back hard on the allegation that it laundered copyrighted recordings through AI training. The filing argues that transformative AI training constitutes fair use under U.S. copyright law and challenges the labels to demonstrate actual copying or market substitution — not just the theoretical risk of it. As the first major AI music copyright cases move from complaint to contested litigation, the legal arguments crystallizing in this case will likely set the boundaries for AI and creative work for years.

What the Lawsuit Alleges

UMG and Sony's second lawsuit against Suno alleges that the company trained its music generation model on commercially released recordings without authorization, effectively laundering those works into AI outputs that compete with the original music in the streaming and licensing markets. The labels' theory is that even if Suno's outputs don't sound exactly like specific songs, the process of training on protected works without a license constitutes infringement — and that the downstream AI outputs derived from that training are themselves infringing.

This is the same core theory behind the labels' first lawsuit, which Suno has been contesting since 2025. The second lawsuit expands the scope of alleged infringement and adds Sony Music as a co-plaintiff alongside UMG.

What Suno's Answer Argues

Suno's 44-page response makes three central arguments:

1. AI training is transformative use. Suno argues that training an AI model on copyrighted works is categorically different from copying those works for distribution. The model learns statistical patterns — relationships between musical elements — rather than copying the works themselves. Under the transformative use doctrine in U.S. copyright law, this kind of learning is protected as fair use.

2. No direct copying, no infringement. The filing challenges UMG and Sony to identify specific instances where Suno's outputs reproduce protected expression — melodies, lyrics, distinctive arrangements — from specific copyrighted recordings. Without that direct copying evidence, Suno argues, the complaint rests on speculation about training methodology rather than demonstrable harm.

3. Market substitution is not proven. Copyright's fair use analysis requires courts to consider whether the allegedly infringing work harms the market for the original. Suno argues that AI-generated music occupies a different market segment from commercially released recordings — that someone using Suno to generate background music for a YouTube video is not displacing a licensed Sony track in a streaming playlist.

Why This Case Matters Beyond Music

The legal questions in Suno v. UMG/Sony are not limited to music. The same doctrinal arguments — about whether training on copyrighted works constitutes fair use, and whether AI outputs derived from that training infringe the originals — are at the center of pending copyright cases against Anthropic, OpenAI, Stability AI, and other generative AI companies.

The music cases have an advantage for establishing precedent: the harm is more measurable. Courts can compare AI-generated music to source recordings using acoustic analysis in ways that are harder to apply to text or image copyright claims. A ruling on transformative training in a music case will be cited in every subsequent AI copyright dispute across creative media.

What the labels want: A ruling that AI training on copyrighted works requires a license — which would force every major AI company to renegotiate training data rights with rights holders, potentially restructuring how AI models are built and who controls that process.

What Suno wants: A ruling that transformative AI training is fair use — which would eliminate the licensing obligation and establish a legal safe harbor for the existing training practices of virtually every AI company.

The Stakes for Creators

Independent musicians and creators are watching this case with conflicting interests. Many support the labels' position — they did not consent to having their music used to train a commercial product that now competes with them. Others, particularly those using AI tools in their own creative process, are concerned that a broad ruling requiring training licenses would effectively halt the development of AI music tools they rely on.

The case is also shaping commercial negotiations that are not waiting for the courts. Several streaming platforms, publishing companies, and AI music tool developers have begun exploring licensing frameworks for AI training — essentially hedging against a ruling that might make unlicensed training retroactively infringing.

What to Watch

Suno's 44-page answer now sets up the discovery phase, where the labels will push for access to Suno's training data logs, model architecture details, and internal communications about training methodology. If discovery reveals direct copying — recordings stored in training datasets, used verbatim rather than learned from — the fair use argument weakens significantly. If Suno's pipeline proves to be genuinely transformative at the technical level, the fair use case strengthens. The outcome of discovery will likely determine whether this case settles or proceeds to trial.