SubmitHub study finds AI signs in 40% of July's new music

A new study from SubmitHub found that almost 40% of music released globally in July showed signs of AI involvement in the production process. The company's AI music detector analyzed more than a million releases from the month, finding that 23.2% were fully AI-generated and another 15.3% showed signs of AI-generated audio that had been "modified or processed" by humans.

SubmitHub, a platform that lets artists submit music to playlists and blogs, unveiled its detection tool, SH Labs, in June with the goal of addressing what it called a "growing disclosure problem" in the industry. "I think the path forward for AI music is disclosure," SubmitHub founder Jason Grishkoff said. "People should be able to decide for themselves whether they want to engage with AI-generated music." Grishkoff added that the company's job is simply to give listeners enough information to make that choice themselves.

The study also found a gap between what artists say and what the tool detects: 31% of artists whose music was flagged as containing AI this year told SubmitHub, when asked, that they had not used any AI tools at all.

SH Labs is already being used by platforms including Bandcamp and Traxsource to flag AI-generated submissions, alongside performance rights organisation GEMA, which won a landmark German copyright case against Suno in July. That ruling found Suno was not entitled to use music represented by GEMA to train its models, and that the company had breached both German and US copyright law.

SubmitHub has not said whether it plans to publish similar monthly breakdowns going forward. Beatport separately moved this month to ban AI-generated tracks outright, a sign the disclosure debate SubmitHub is tracking is starting to shape platform policy as well.