AI in the Music Industry – Part 16: A Neighbouring Right for AI-Output?

As we saw in part 15 of the series ‘AI in the Music Industry’, artificial intelligence will not have a copyright personality for the foreseeable future. However, it is being discussed that the output created by AI creates could be protected by a related or neighbouring right, as in the case of images (film and photography) or sound recordings (neighbouring right for music labels). In this part of the series, we look at the advantages and disadvantages of a neighbouring right for AI-generated musical works.

Unlike copyright, neighbouring rights are not linked to the act of creation, but to the result of the performance, e.g. the music recording. Since music AIs are also capable of producing recordings, this would certainly be an interesting legal starting point. However, the question remains who could be benefit from such neighbouring rights. Is it the creator of the AI software or its provider or the person who formulates the prompts to the AI? In any case, Universal Music Group is certain that users of an AI can in no way establish authorship of the output created by the AI because “(…) authorship requires human expression, and the output of generative AI systems does not embody human expression.”[1] However, the prompts to the AI could be copyrighted if they reach a certain level of originality and creativity, according to Universal.[2]

But it’s not that simple. There are already laws in place, such as the UK Copyright Designs and Patent Act of 1988, which states that literary, dramatic, musical or artistic works created by a computer should be protected, with authorship vested in the person who makes arrangements or preparations of such a nature as to enable the computer to create the work.[3] Of course, the computer is seen here as an aid to the creation of a work, but an analogy could certainly be drawn with the operation of an AI system, which would then be entitled to a related property right in the work produced.

Another option would be to protect AI services by competition law, as proposed by the German Industry 4.0 platform, if whether the AI output has a competitive character from which the company’s specificity can be inferred.[4] However, such a related right would be much weaker than a copyright-related property right and would only provide protection against imitation but would not entail any remuneration obligations.

Therefore, Platform Industry 4.0 is in favour to establish an ‘AI neighbouring right’: ” The creation of a separate neighbouring right would have the advantage of being able to adequately reflect the special features of ‘AI creation’ and centrally regulate issues such as the ownership of the rights to these results” [translation by the author].[5]

However, there are also massive reservations about such ancillary copyright for AI systems. The former Chief Technology Officer (CTO) of the Recording Industry’s Association of America (RIAA), David Hughes, posed the following thought experiment in an interview: “If I were, say, a ‘copyright anarchist’ and wanted to maximise profit, I would use AI to start creating every mathematically possible composition that is reasonable or likely to exist in the future based on samples of as many commercially successful sound recordings as possible. I would plug into a cloud service like Google and create billions of sound recordings in an easily indexable system. Then, when the next Nicki Ninaj hit comes out, I go through my database and I look, and I see, oh, wait a minute, those few bars match up with the few bars that I wrote last year. Now I have got a potential copyright claim.”[6]

Even if the technical feasibility of this project can be questioned, the thought experiment points to a fundamental problem with the legal protection of independently created works by an AI. Indeed, such protection could be used as a barrier to entry. A generative AI can produce vast quantities of musical works and recordings in a very short time, which, if all protected by copyright, could be used to defend against music created by humans. This, in turn, would argue for maintaining the legal status quo with respect to AI music creation and not granting related rights to AI musical works and recordings.

---

Endnotes

[1] Universal Music Group, 2023, Notice of Inquiry “Artificial Intelligence and Copyright” before the United States Copyright Office, Docket No. 2023-6, p 75.

[2] Ibid.

[3] Sec 9 (3) of the United Kingdom Copyright Designs and Patent Acts points out: “In the case of a literary, dramatic, musical or artistic work which is computer-generated, the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken.”

[4] Plattform Industrie 4.0, 2019, “Künstliche Intelligenz und Recht im Kontext von Industrie 4.0”, April 2019, Berlin: Bundesministerium für Wirtschaft und Energie, p 26.

[5] Ibid., p 27.

[6] Cited in Martin Clancy (ed.), 2023, Artificial Intelligence and Music Ecosystem, London & New York: Routledge, pp 104-105.