AI music generators sued over use of copyrighted songs in training
Major record labels and music creators are suing AI music generator companies for training their systems on millions of copyrighted songs without permission or compensation. The dispute centers on whether using copyrighted music to train AI models violates intellectual property law and could reshape copyright enforcement in the digital age.
What happened
AI music generation systems were trained on databases containing millions of copyrighted songs without the permission of artists or rights holders. According to an investigation by journalist Alex Reisner, the datasets used to develop these systems contained between hundreds of thousands and over 12 million songs spanning nearly all musical genres, including links to songs on platforms like Spotify and YouTube.
Musicians, composers and major record labels argue that using their copyrighted works to train algorithms without consent or compensation constitutes a serious violation of intellectual property law. Record labels including Warner, Sony and Universal have filed lawsuits against the developers since 2024.
Reported impact
The music industry contends that AI-generated melodies often reproduce stylistic and structural elements similar to existing copyrighted pieces, allowing companies to benefit commercially from artists' work without providing any compensation to the original creators.
Index analysis
The case involves questions about whether training AI systems on copyrighted material without permission falls within fair use or constitutes infringement under copyright law. The outcome could fundamentally change how copyright rules are applied in the digital age and AI era.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.