Sony's AI Music Attribution Tool Targets Copyright in Generative AI

Sony AI, the research division of Sony Group, has developed a technology capable of identifying specific musical influences in AI-generated content through a method called "unlearning," according to a research paper published on Sony's official AI website.

The technology, detailed in a NeurIPS 2025-accepted paper titled "Automatic Music Sample Identification with Multi-Track Contrastive Learning," works by selectively removing a generated track from an AI model and measuring which training songs are most affected by that removal. This approach allows the system to pinpoint which musical works most influenced an AI output, even when the output doesn't obviously resemble the source material.

As demonstrated in the research, the system can estimate specific influence percentages, such as determining that an AI-generated track contains "30% influence from the Beatles and 10% from Queen." The accuracy of these identifications depends on whether AI developers cooperate with Sony by providing direct access to their model's training data.

When developers cooperate, Sony can obtain data directly from the base model system for appropriate licensing and compensation. Without cooperation, the system compares AI-generated songs to existing music catalogs to identify original works. Sony AI describes this research as part of its broader mission to explore "musical integrity in the age of machine learning, exploring attribution, recognition, and protection."

This development positions Sony within an emerging ecosystem of copyright management tools for generative AI, joining companies like Sureel, Musical AI, and ProRata. While Sony's technology is still in research and development, it represents a significant step toward creating systems that could distribute revenue generated by AI music to original songwriters based on their contribution.

Industry analysts note this approach differs from traditional copyright detection systems, which typically identify exact matches rather than influence percentages. The technology could help resolve ongoing disputes between AI developers and music rights holders, particularly as platforms like Suno have recently reached licensing agreements with major labels like Warner Music Group.

Sony AI, established in April 2020 as a division of Sony Group, focuses on "pursuing groundbreaking research in AI and robotics to unleash human imagination and creativity with AI." The company has published three papers accepted at major academic conferences in 2025, with this music attribution research forming part of its broader effort to address copyright challenges in AI-generated content.

Artūras Malašauskas is an AI Systems Integrator with 20+ years of production-grade web engineering experience. He has designed, shipped, and scaled enterprise Python/PHP systems for logistics, SaaS, and public-sector clients. For the past year, he has focused exclusively on AI integrations: deploying open-source LLMs, building generative media pipelines (image, audio, video), and engineering multi-agent workflows for real production environments. His standard: reproducibility, security, cost-efficient inference—no vaporware. He documents and evaluates emerging AI tooling, separating verified capabilities from marketing noise. Technical editor at: muza-ai.eu, ai-verslas.lt, ai-naujinos.lt Connect on LinkedIn