AI music generation traced to source audio for verifying contribution
Tracing Inputs, Verifying Outputs: Validating Attribution in Music Generation
SoundMachine Learning
Summary
It can be hard to tell which music was used to create new AI-generated songs. The authors show a way to track the original audio inputs that influenced the AI’s output and check how much they shaped the result. They built tools to see if generated music copies training data not used as input, finding little such copying. This helps prove whose music really contributed, which is important for fair rights and payments as AI music grows.
What this means in practice
- •For music producers: Check which original audio sources influenced AI-generated music to manage usage rights and compensation accurately.$Commercial implications: Enables commercial rights management and licensing tools to verify music contributions and protect copyright in AI productions.
- •For audio software developers: Integrate input attribution methods to provide users transparency on how AI music generation used their provided audio.
Authors
Taejun Kim, Wonil Kim, Jongmin Jung, Hyeongseok Wi, Sangeun Kum, Keunhyoung Luke Kim, Taehyoung Kim, Dongjoo Moon, Seungsoon Park, Taewan Kim, Virginie Berger, Juhan Nam, Jongpil Lee
Abstract
How can we verify whose music contributed to an AI-generated output? This paper demonstrates how input-based attribution can provide verifiable evidence of which audio sources were used in a generation and whether they shaped the output. To do so, we condition the generation solely on audio without any text input, then trace the inputs behind each output, and establish their musical effect. In prompt adherence tests and controlled input swaps, the stems generated by our generator, MixAudio, follow the prompt audio in timbre and the context audio in harmony. Yet these outputs may still reproduce training data not supplied as inputs. We therefore audit memorization with our musical version identification model, musicDNA, and find few reproductions outside the input records. On human-judged cases within the flagged pool, it achieves higher precision and recall than the other tested memorization detectors. The two evaluations suggest that input records and output analysis provide complementary evidence for attribution, on which rights-holder reporting and compensation can draw as the AI music economy takes shape. Audio examples are available at https://neutune.github.io/attr2027demo/