Introduction

Artificial Intelligence (AI) is no longer a futuristic concept in the music industry or any other sector. For at least the past two years, AI has been used to produce songs, generate and modify vocals, and create lyrics. Using platforms like Suno and Udio, entire albums can be generated in minutes. This implies that the use of AI in the music industry is no longer experimental. AI songs are already available on Spotify, Apple Music, and other digital streaming platforms, generating revenue while providing entertainment for fans. The question is no longer whether AI will affect music. In fact, it already has. The real question is how the law will respond, and how creators can protect themselves. This article will explain the meaning of AI-generated and AI-adapted music, the legal risks involved, the ethical considerations raised, and the questions that creators and businesses should consider.

What is Meant by "AI Music"?

AI music can generally be categorized into three types: music created entirely by AI systems; music created by humans with AI assistance; and music that uses AI to adapt, remix, or clone existing works. This distinction is crucial because the law responds differently to each category. Each category is discussed below.

Music Entirely Generated by AI

This category refers to music produced with little or no human creative input. A prime example is an AI user merely inputting the prompt "Create a sad, Burna Boy-style Afrobeat song" into Suno, and the AI generates a complete song from scratch, mimicking Burna Boy’s style, including lyrics, melody, and vocals. The user does nothing but input the prompt. Human creative contribution is minimal, or entirely non-existent.

A real-life example is "Breaking Rust"—an AI-generated country music project. Its song Walk My Walk topped the Billboard Country Digital Song Sales chart in November 2025. This means the song generated significant revenue for the operators behind the project. However, if a system creates a song almost entirely on its own based solely on a prompt, one must naturally begin to question authorship and ownership.

Copyright law assumes that the author is human, and creative activity must involve human judgment. The law will not recognize any "non-human" output as an author. Looking back at the US "Monkey Selfie Case" in 2018, the court ruled that a photograph taken by a monkey using a camera could not be protected by copyright. Since AI is not human, it certainly cannot be an author or creator.

So, who should be recognized as the author of music entirely generated by AI? This is a vital question. Without an author, copyright cannot be established. A common view is that an AI user who only inputs prompts cannot be an author. The opinion of the US Copyright Office is widely adopted: "Prompts alone do not provide sufficient human control to make the user of an AI system the author of the output. Prompts essentially act as instructions, conveying ideas that cannot be protected." Merely inputting a prompt does not allow a person to determine the result in a substantive way. The prompter only has a vague expectation of what they consider a good result, while the final result itself is fortuitous.

Nor can one recognize AI developers as authors. The developers' contribution is not creative in nature. It can be compared to merely providing a recording studio, which is insufficient to confer authorship rights.

If copyright law struggles to identify any party as an author due to a lack of meaningful human creativity, this may mean that music entirely generated by AI belongs to the public domain. This implies that anyone is free to copy, distribute, or adapt such music without permission. For enterprises investing in AI-generated content, this presents a serious commercial risk, as they may not actually own what they believe they own.

Music Collaboratively Generated by Humans and AI

In these scenarios, people use AI tools to assist in the creative process, rather than letting AI do all the work. This includes instances where an artist writes the lyrics but uses AI to suggest the melody, or a producer creates a beat but uses AI to optimize the arrangement. A good example is the song I Run by HAVEN, where the creator, Harrison Walker, wrote the song himself but used Suno to process his vocals to make them sound like a female singer.

From a legal perspective, this does not undermine authorship. Since the human creator retains control and makes meaningful creative decisions, the output remains protected by copyright. This is equivalent to using AI as a tool, no different in principle from using Auto-Tune, drum machines, or Digital Audio Workstations (DAWs). If using Auto-Tune does not strip authorship, then using AI solely as a tool should not either.

However, the final result still depends on the degree of AI vs. human contribution. If AI makes most of the expressive decisions, the justification for protection weakens. If a human curates, edits, and directs the result, the case for protection is stronger. Of course, this is not a matter that can be determined with mathematical precision; it must be judged on a case-by-case basis.

It should be noted that even if a human retains authorship, using AI as an auxiliary tool still carries other legal risks. AI systems are trained on music created by real artists, and if the output begins to mimic a recognizable voice or performance style, issues regarding privacy and publicity rights may arise. For example, the vocals in HAVEN’s I Run sounded very similar to Jorja Smith, leading to controversy and industry pushback, as Jorja Smith’s record label claimed the track used AI to mimic or "clone" her voice, misleading listeners into thinking she sang the song.

Existing Music Adapted Using AI

In this category, legal risks increase significantly. This may include AI remixes of popular songs or AI cover versions generated using cloned voices. A typical example is the AI version of Stromae’s original song Papaoutai, titled Papaoutai Afro Soul. In these cases, there is an existing work protected by copyright that has been adapted or remixed by AI.

AI adaptation may trigger infringement on multiple levels, including copyright in the musical work and/or sound recording, performer’s rights, and moral rights (e.g., distortion or derogatory treatment of a work). If the output is clearly identifiable as originating from an existing work, a license must be obtained. Failure to obtain a license could expose creators, streaming platforms, and distributors to infringement claims. Since the release of the AI version of Papaoutai, Stromae has remained silent. On a non-public level, he has likely licensed the remix or reached some favorable royalty arrangement. If not, the possibility of an infringement claim remains.

Training Data Issues

Although AI systems learn by analyzing vast amounts of existing music scraped from the internet, these songs were generally not used with the consent of the creators. Many artists never agreed to have their work used to train the very AI machines that now compete with them. This is also one of the key points where law and ethics intersect. Whether it is legal to use copyrighted materials to train AI remains unsettled globally. The current debate focuses on whether such training acts constitute copyright infringement or fall under fair use or fair dealing. The main argument supporting fair use is that the usage is transformative. Proponents argue that training AI to use copyrighted works is not for the purpose of copying or distributing those works, but to analyze the patterns, styles, and structures within them to generate new output, which makes the usage fair. However, this conclusion is doubtful.

Courts generally consider four factors when determining whether a usage is fair: the purpose and character of the use; the nature of the copyrighted work; the amount and substantiality of the portion used in relation to the copyrighted work as a whole; and the effect of the use upon the potential market for or value of the copyrighted work. The argument for transformative use relies solely on the first factor, ignoring the others. In Kadrey v. Meta Platforms, Inc., experts noted: "There is absolutely no rule that says if your use of a protected work is transformative, it automatically immunizes you from a claim of copyright infringement." If an artist's entire catalog is used to train AI systems that now compete with or even replace them in the market, it is difficult to call that truly fair.

Ethical Issues

Beyond legal issues, the rise of AI in the music sector presents critical ethical concerns for creators and the industry. These questions touch upon the core of fairness, creative integrity, and respect for the individual identity and labor of human creators.

Creative Integrity

AI raises a series of important questions regarding creative integrity: Does it weaken human creativity? Or does it expand it? Where does originality stand now? Platforms like YouTube allow AI-assisted content to be monetized, but they increasingly require transparency regarding synthetic media and impose rules on originality. Legal or not, creators must consider whether using AI aligns with their artistic identity and brand positioning.

Economic Impact

AI has changed the economic model of music. It reduces production costs, increases supply, and makes the market more easily saturated. A major consequence is the reduction of licensing opportunities. If businesses can instantly generate "good enough" music, they may no longer need to license existing songs or commission composers. This reduces income sources for "original" creators. As AI-generated music becomes abundant, "original" creators may become easier to overlook. This economic pressure is real, especially for independent artists and composers, raising the question: Do people really want to support AI to the fullest extent?

Identity and Personality Issues

AI can replicate voices, accents, and performance styles. This goes beyond the scope of copyright and touches on individual identity, reputation, and personal dignity. Assuming a mimicry does not infringe copyright, is it ethical to release a song with someone else's voice without their consent?

Globally, regulators are issuing guidance, debating specific legislation for AI, and responding to court rulings. However, the pace of technological development always seems faster.

Practical Advice for Creators and Businesses

Waiting for relevant laws to catch up is risky. Before laws and court rulings are established, ownership and usage rights should be determined by contracts and terms and conditions. For example, if an independent artist or a record label distributing music explicitly states in their terms and conditions that their songs may not be used to train AI, then violating this provision is not only a matter of copyright infringement but also a breach of contract. Contractual protection is vital.

Furthermore, when enforcing rights (e.g., when an artist’s work is used to train AI), suspicion is not enough. To obtain relief for infringement, the party must first prove that their work was used, and that the usage constitutes infringement. This is certainly not a matter to be settled by sentiment. Evidence is crucial and must be provided. This appears to be a major challenge facing plaintiffs in many AI-related lawsuits.

To stand on firm ground both ethically and legally, creators should document their creative processes, obtain consent for the use of voices and likenesses, avoid mimicking recognizable artists, and explicitly define how AI may be used in recording, publishing, and production contracts. Importantly, creators should continue to develop truly creative works. Although AI can provide assistance, originality remains a valuable asset.

Conclusion

AI-generated and AI-adapted music brings both opportunities and risks. It challenges traditional notions of authorship, ownership, fairness, and individual identity. For clients operating in the music sector (whether as creators, investors, or distributors), the key is to raise awareness and engage in proactive planning. AI is not going away, and the legal and ethical responsibilities that come with creation will not go away either. Like it or not, the future of music will involve the participation of both humans and machines. The task at hand is to ensure that innovation does not come at the expense of fairness, integrity, and respect for originality.