AI Music Digest

The Australian Recording Industry Association has updated its Charts Code of Practice to exclude recordings made wholly or mostly with generative AI. Under the revised code, a recording built with generative AI is eligible only if it is “substantially human-made” and raises no stream or chart manipulation concerns. ARIA says it will apply the definitions in the labelling standard published by the global music community on July 10.

The change takes effect with the ARIA Chart dated Monday, August 31, published Friday, August 28. AI used in a supporting production role remains eligible; recordings with AI-generated lead vocals or primary instrumental performance would not be. ARIA retains authority to remove ineligible recordings, adjust chart positions and revoke awards, with a process for artists to contest exclusions.

The move follows scrutiny of Queensland producer Josh Fawaz’s cover of Madonna’s “Like a Prayer,” which reached the top of Australian radio airplay in July and peaked at number four on two ARIA charts. Fawaz later disclosed on the track’s Spotify listing that the vocals and drums were AI-generated. ARIA CEO Annabelle Herd said the changes “reflect our intent to remain dynamic and promote the human nature of artistry.”

Why It Matters: ARIA is the first national chart to operationalize the IFPI-backed eligibility principles published in July. It converts an industry statement of intent into an enforceable rule with a defined appeals process — the template other chart bodies are expected to copy.

Stability AI announced on August 25 that it has closed a $76 million Series B whose investors include all three major music companies — Universal Music Group, Sony Music Group and Warner Music Group — alongside Electronic Arts, AMD Ventures and Pacific Alliance Ventures. Existing backers Coatue, Greycroft, Kadmos Capital, Sean Parker and Eric Schmidt participated for a second consecutive round. Variety reported the round first, as an exclusive.

The company says the round brings its total raised under CEO Prem Akkaraju to $232 million and that the capital will go toward its creative production product suite, applied research and professional services arm. Stability describes itself as the first AI company to take direct equity investment from all three majors in a single funding round — a claim attributed to the company and its backers rather than independently verified.

The investment builds on licensing and model-building partnerships Stability struck with UMG and Warner in late 2025, which underpinned Stable Audio 3.0, the licensed-data model family it released in May.

Why It Matters: The majors have moved from litigation to licensing to ownership. Equity gives labels upside and governance leverage over a generative model rather than only a royalty stream — while raising unresolved questions about whether artists share in that upside.

Billboard said on August 25 that it will work toward greater transparency around AI in music, including identifying songs that use AI on its charts and in its editorial coverage. Billboard framed the commitment as depending on two things: creator and content-owner disclosure, and reliable detection tooling. It did not announce a ban or a change to chart eligibility.

Chart data provider Luminate announced a companion framework the same day. Luminate says its CONNECT platform will analyze streaming data, metadata and audio to flag AI-generated songs and artists, applying an “AI Generated” label to songs and artist profiles, with a review process for artists and labels to challenge incorrect labels. The company says it is also working with major DSPs and AI music companies to pull self-reported streaming and attribution data into CONNECT. The feature is slated to reach customers and partners later this year.

Why It Matters: Billboard has had no formal AI chart policy while other chart bodies moved. Labeling rather than exclusion is a softer line than ARIA’s, but Luminate’s data layer is the piece that would make any future US eligibility rule enforceable.

Independent publisher Round Hill Music filed separate copyright complaints against Suno and Anthropic in the US District Court for the Northern District of California on August 17. The suits allege both companies used works Round Hill owns or administers to train AI models without license or compensation, and that they used scraping software to bypass technical protections in violation of the DMCA.

The complaints name works including the Goo Goo Dolls’ “Iris,” Bonnie Tyler’s “Total Eclipse of the Heart” and James Brown’s “I Got You (I Feel Good).” Round Hill listed roughly 500 songs but said it will amend to cover at least 10,000, arguing damages would run into the hundreds of millions and “may ultimately approach or exceed $1 billion” in each case. Round Hill has demanded a jury trial and, per Music Business Worldwide’s reporting, has signaled it is not looking to settle or license.

Why It Matters: Most publishers and labels that sued AI firms converted the claims into licensing deals. A publisher explicitly pursuing a verdict instead keeps the underlying fair-use question live for the catalog owners who have already settled.

Music IP Holdings, a venture formed from a partnership between Universal Music Group and IP management firm Liquidax, has licensed a portfolio of more than 24 patents to AI music platforms Udio and GRAI. The company says more than 50 further patents are pending, and that the portfolio is available to license through an online platform.

According to the announcement, the patents cover moderation, watermarking, licensed distribution and monetization, and extend beyond music to audio, video, film, animation, images and text, including name-and-likeness applications. Music IP Holdings says the framework is intended to “drive responsible innovation while providing additional ways to protect artists, songwriters and their music,” and that the licenses will support development of licensed AI services including covers, remixes and interactive music experiences. Variety reported the deals first.

Why It Matters: Copyright litigation contests how models are trained. Patents govern how licensed AI music products are built and monetized — a second, quieter control point where a major label is positioned to collect from platforms regardless of how the training cases resolve.