International music publishing association claims it has “extensive evidence of serious copyright infringement” by AI firms that have used popular songs to train generative modelsAn international music publishing company called ICMP claims it has gathered two years worth of evidence showing that AI firms have scraped copyright-protected music from millions of artists to train their generative AI models.
Similar claims have already been brought to lawsuits against firms like Udio and Suno by major labels (namely one in 2024), and even by an independent artist this year, but ICMP’s findings appear to be the largest investigation so far into the training of generative AI across the music industry.

READ MORE: AI finally pays up: Beatoven.ai’s new Maestro AI music generation model promises royalties for artists

Its findings have been shared exclusively with Billboard, which reports that the investigation was carried out using publicly available registries, open-source repositories of training content, leaked materials, research papers, and independent research by AI experts.
ICMP claims to have compiled “comprehensive and clear” evidence of the unlicensed use of digital music for AI training and GenAI music, as well as songwriter and performer image outputs. ICMP also says that the scope of the training is larger than previously acknowledged.
Documents shared with Billboard name a large number of AI firms including Suno and Udio, as well as Microsoft’s AI app CoPilot, Google’s AI system Gemini, OpenAI (owner of ChatGPT), Midjourney, and more.
According to the report, an “admission” from OpenAI’s chatbot (following enquiries from ICMP) allegedly revealed that its OpenAI Jukebox music-making app was trained on music by artists including The Beatles, Elton John, Madonna, Beyoncé, and many others.
The report acknowledges that when OpenAI launched Jukebox in 2020, the company publicly disclosed that it had trained the app on a dataset of 1.2 million songs, though it never revealed what songs or artists were used. ICMP’s findings also claim that X’s chatbot Grok is “one of the worst offenders when it comes to respect for songwriters’ and artists’ rights”.
“This is the largest IP theft in human history. That’s not hyperbole. We are seeing tens of millions of works being infringed daily,” says ICMP director general, John Phelan. “Within any one model training data set, you’re often talking about tens of millions of musical works often gained from individual YouTube, Spotify, and GitHub URLs, which are being collated in direct breach of the rights of music publishers and their songwriter partners.”
Phelan continues, “Despite their public claims that they’re not training upon copyright-protected works, we’ve caught many [tech companies] red-handed. We have extensive evidence of serious copyright infringement. Many of these companies are scraping the lyric datasets from the internet of millions of works and putting them into their models. Aside from amounting to breaches of copyright laws and often contract laws, this is often done despite the music sector’s consistent and clear statements that licenses are both required and available for legal AI training and GenAI.”
In previous complaints and lawsuits brought against AI firms, many have used the defence of “fair use”, which permits the limited use of copyrighted material without need for permission from the rightsholder(s). Suno has previously claimed that “what the major record labels really don’t want is competition”.
Billboard says it has contacted all the tech companies mentioned by ICMP, and that all of them either declined to comment or did not respond to its requests.
The post International music publishing association claims it has “extensive evidence of serious copyright infringement” by AI firms that have used popular songs to train generative models appeared first on MusicTech.