In early 2025, Norwegian blogger Andy Hafell, known as AI Andy, used the music-generating neural network from Suno to create an album of instrumental lo-fi hip-hop tracks. The result was a sort of “elevator music”—unobtrusive background tunes that fall into the category of “study/work vibes.”
Andy then utilized distributor Distrokid for publishing his album on streaming platforms. He noted that during the release process he had no option to specify that the music was generated by artificial intelligence.
The album — released under the alias "Haffy." (and no, that's not a typo; the period is actually part of the name) — appeared on all major streaming services and remained available there through mid-May 2025. According to the blogger, he provided his real name to the distributor. It seems nobody even bothered to verify whether this release involved any AI-generated content.
Some musician-bloggers take things further. Beyond English-speaking YouTube, one can find video segments titled something like “AI Wrote My Song, and No One Even Noticed.” These artists craft lyrics, upload them to generative systems like Suno or Udio, generate relatively engaging tracks, and then recreate those tracks in their own studios using live instruments, synthetic sounds, samples, and their own vocals. They proceed to release these works just as AI Andy did.
This kind of activity raises serious ethical questions. Even if bloggers and artist-bloggers intended to draw attention to the issue, they often don’t remove their releases from streaming platforms after posting their videos. Why might that be?...
Unfortunately, behind these publicized cases lies much more unpublicized cheating in the music industry. Behind the scenes, musicians worldwide discuss instances where melodies, harmonies, hooks, and even stylistic elements are first created using neural networks before being recreated in-studio by cheater-artists and published under their names.
Anyone looking to cut corners in this way, including those doing so openly, should know that co-writing with AI and then entering the professional music scene falls into a zone of maximum legal uncertainty—especially in the U.S. market. The United States is the primary testing ground where lobbyists for AI companies and advocates for human creators are locked in a tug-of-war over how the law defines music authorship. Decisions made in the U.S. often shape global standards. Other countries tend to adopt similar definitions soon after, which means that the outcome of these legal battles will likely echo far beyond American borders.
On May 9, 2025, the US Copyright Office publicly released a pre-publication version of its new report on copyright and artificial intelligence. Among other points, it states (direct quote):
Less than 24 hours after the document's revised edition went public, Shira Perlmutter, head of the US Register Of Copyrights at the Library of Congress whose team authored the report, received notice of her dismissal. This development caused significant alarm among every society, association, and publisher representing human musicians who create music through their own labor and talent.
Analysts link this firing to allegations that Shira Perlmutter may have been aligned with music industry lobbyists, having attended events organized by groups like the National Music Publishers Association (NMPA) over recent years.
Meanwhile, President Trump, who effectively terminated the country's top official responsible for copyright issues, surrounds himself with numerous lobbyists from AI-service companies, including musical ones. Some of these entities could soon face court proceedings alleging that they trained their neural networks on publicly accessible music catalogues without obtaining proper licenses.
Put simply: the struggle between people in the music business who want to protect the rights of human creators and those in the AI business aiming to strip away those protections has escalated to dramatic levels.
What’s at stake here is that beneficiaries of AI services intend to prove in the near future that unauthorized use of human-made music—and, by extension, any form of human-produced art—is legitimate, ethical, and acceptable.
This battle unfolds amidst ongoing litigation efforts initiated by several large holders of music rights in the US, both individually and via the Recording Industry Association of America (RIAA). Plaintiffs aim to demonstrate that utilizing copyright-protected materials for training AI models without permission constitutes illegal behavior. And they stand a good chance of succeeding.
If big tech giants lose in court, licenses issued by AI music services for various uses of generated songs and instrumentals would become worthless in traditional music and publishing industries. AI-generated music would be relegated to mere entertainment toys for enthusiasts.
The lawsuits filed in 2024 against startups Suno and Udio by major record labels, including Universal Music Group, Warner Music Group, and Sony Music, carry substantial evidence. The plaintiffs allege that these companies used copyright-protected recordings to train their AI models without authorization from rightsholders. As a result, the AI systems generated tracks similar to iconic songs such as Chuck Berry's "Johnny B. Goode" and Mariah Carey's "All I Want for Christmas Is You."
One example involves Suno's generation of a song nearly identical in melody and lyrics to Berry's famous hit, complete with the distinctive chorus line "Go, Johnny, go."
Of course, these legal battles will drag on for some time. However, their very existence already poses a considerable risk for artists and composers experimenting with co-writing using AI. Labels, particularly publishers specializing in production music for TV, trailers, and advertising, will scrutinize each track mentioned to involve AI usage with a magnifying glass.
Similarly, this scrutiny will apply broadly to any synchronization of music in films, documentaries, games, and other forms of audiovisual media.
Why is this important? Because “cleared rights” for music placed in affordable productions as well as expensive synthetically-generated works like Hollywood movies are considered sacred cows legally by music publishers, content producers, studio legal departments, and everyone else involved in production.
Currently, in the United States, works generated entirely by AI do not receive copyright protection. In other words, they cannot legally have an author. However, when AI generates tracks eerily similar to compositions originally created by humans, a paradoxical collision arises: the authors of these seemingly AI-created tracks might turn out to be either deceased individuals or living persons.
Even if you're an experimental artist who doesn't plan to engage a sync agent or submit your work to production music libraries, you could still face unexpected claims from copyright holders of previously released music down the road. This is because your hybrid human-AI track might resemble someone else's composition too closely.
If you're experimenting with co-writing alongside AI, you're currently in a legal "superposition." On one hand, you might contribute significantly more to the creation of a piece compared to the AI's share. On the other hand, questions about authorship of hybrid creations remain largely unresolved under US law and seem to lack clarity in most global jurisdictions.
Should serious concerns arise regarding your track, you'll need to demonstrate that it:
As expected, proving these points can potentially become lengthy, painful, costly, and possibly scandalous.
Let's consider the service AIVA as an example. Unlike the aforementioned startups, AIVA focuses not on producing ready-to-use audio files but rather on assisting in composing music. Specifically, AIVA delivers MIDI files sequentially, allowing the composer/producer to refine the music independently in their Digital Audio Workstation (DAW).
However, copyright ownership of the final product is granted only to users subscribed to the Pro Plan. At first glance, everything appears perfect: the platform asserts that users can monetize jointly created tracks "in ANY way," quoting directly. Additionally, master recording rights will belong to you since you’ll create them yourself…
But delve deeper into the terms of service, and you'll discover a limitation: monetization is permitted under the Pro Plan conditions only while the composer/producer remains an individual or runs a small business employing fewer than three employees. Moreover, the individual or small business must earn less than $300,000 annually.
If fortune smiles upon you and you become part of something bigger as an artist or composer, you'll need to approach AIVA to negotiate custom terms with the service.
To sum up, there's currently no legally safe and transparent method for co-writing original music together with AI. All existing approaches and services come with risks, sometimes critical ones.
Turning to discussions on co-writing with AI within specialized online communities involving music publishers—for instance, trailer, production, and advertisement music—the consensus among professionals is clear: they won't touch AI-generated music.
The music industry watches anxiously as legal disputes unfold against Suno and Udio. Globally, collective professional initiatives emerge aimed at protecting the rights of flesh-and-blood music creators (including their right to work). If humanity loses to lobbying efforts supporting widespread parasitic exploitation of generations-old artistic achievements by AI, profound changes—or perhaps total transformation—are inevitable. A select few (the minority) stand to gain substantially, whereas professionals (the majority) will suffer greatly.
A superficial pop-music blog might end this article with a checklist titled "What to Consider When Co-Writing Music with AI for Legal Monetization." But we're not going to provide one. After all, it's essential for those who comprehend the amount of sweat, blood, and tears poured into making great music by their peers to support fellow humans.
Is this Luddism? Absolutely not. For the most part, the music community isn’t opposed to technologies incorporating AI into creative processes. These tools could evolve into fascinating and even routine assistants. What professionals truly desire is fairness—a game played according to rules.
