Top 5 AI music news of the week (25th – 31st May 2026)

Here are the Top 5 AI Music News of the week.

1. Musician with Parkinson’s finishes album using AI

Samuel Smith, a 49-year-old London musician, finished his album ‘The Art of Letting Go’ after Parkinson’s took his guitar playing. He hummed melodies into his phone and used Suno and Udio to turn them into demos.

He wrote every lyric and every melody himself. The AI built reference demos, sometimes after 150 attempts, that he handed to live session players so they could hear what was in his head. Grammy-winning producer Matt Rollings and guitarist Julian Lage joined him in Nashville, where Smith played a final guitar duet on the track “Horizon.”

“AI is not replacing anything for me,” Smith said. “It’s unlocking, it’s enabling. It’s allowing me to keep writing.” Most of the AI music debate centers on scraped catalogs and fake artists. Smith’s story runs the other way: a working musician using the same tools to keep doing his job. If your body or your budget ever blocks the work, that’s the use case worth remembering.

2. GEMA vs Suno verdict pushed to July 31

The first major European court ruling on AI music training slipped six weeks. The Munich Regional Court moved its GEMA vs Suno decision from June 12 to July 31, 2026.

The reason is procedural, not a signal about the outcome. The court cited internal administrative grounds. GEMA filed back in early 2025, accusing Suno of training on protected works without a license, and played AI outputs it says closely match Forever Young, Mambo No. 5, and Daddy Cool.

A GEMA win would be the first European ruling that AI platforms need a license to train on copyrighted music. For you, the new date changes nothing. Register your catalog with your PRO now, because when licensed AI deals close, payouts route through registration records and market-share data. Clean metadata is what puts you in line to get paid.

3. Publishers drop vicarious claim against Anthropic

Music publishers walked back one of their biggest legal weapons against Anthropic. Judge Eumi K. Lee approved the voluntary dismissal of their vicarious infringement claim over lyrics used to train Claude.

The retreat traces straight to one ruling. In March, the Supreme Court sided with Cox Communications and narrowed secondary copyright liability, the theory that holds one party responsible for another’s infringement. Concord, UMG, and ABKCO then refiled an amended complaint identical to their January version, minus the vicarious claim.

A contributory claim over torrenting still targets co-founders Dario Amodei and Benjamin Mann, so the case isn’t dead. Why you should care: the Cox precedent is making it harder to sue AI companies for what their users do. If you’re counting on these lawsuits to set the rules for AI training, the ground shifted under them.

4. Tonada launches AI in-store music that skips royalties

A Stockholm startup wants to replace the background music in shops and restaurants with AI and cut performance royalties out of the deal. Tonada launched on May 27 with funding from Last.fm founder Michael Breidenbrücker and customers like Hawaii Poké and Vercel already signed.

The pitch is no PROs, no royalties. Tonada generates original adaptive tracks owned by each brand, then shifts the music in real time based on foot traffic, weather, and time of day. Its homepage names ASCAP, BMI, SESAC, GEMA, and SAMI as the per-location fees it routes around. CEO Juan Manuel Serruya ran engineering at Spotify before this.

Retail and hospitality background music is a quiet but large royalty pool, and library composers have written into it for decades. If Tonada and its peers take even 5% of that pool, the per-writer cut on every distribution shrinks by close to the same amount. The squeeze lands on session and library work, the steady gigs that pay a lot of musicians’ rent. Watch this corner of the business, because it’s where AI is taking real income first.

5. Jazz Is Dead launches a free ‘Played By Humans’ stamp

Jazz Is Dead, the label led by producer Adrian Younge, launched Played By Humans, a free platform that certifies music made by real people. Artists and labels stamp their tracks, and fans can upload any song to check whether a human or AI made it.

The timing tracks with the numbers. 44% of new uploads to Deezer are now AI-generated, and AI acts have charted before listeners knew a machine made the music. Younge calls the badge a mark of trust, not a weapon, and says the project isn’t anti-AI, only pro-transparency.

The stamp means something only if enough artists and fans treat it as a signal, and it’s free, which lowers the bar to try. The catch is the binary. A fully prompted Suno track and a human song with one AI drum fill both fail a plain “human or not” test, even though they’re nothing alike. If you make music with any AI in the chain, watch how these badges define “human” before you lean on one to tell your story.