In one of the many iconic scenes of Ridley Scott’s movie Blade Runner (1982), an artificial “replicant” called Rachel plays piano for the detective that is supposed to hunt her down and “retire” her, making him question his own humanity. More than 40 years later, music created by or with artificial intelligence (“AI”) is omnipresent. Both, creators and users often assume that works generated by AI cannot enjoy copyright protection and, therefore, may be used freely. When it comes to music, this is not the necessarily true. While an AI system cannot be “author” under German copyright law, creators of music can gain exclusive rights to compositions and recordings generated by or with AI technology. Users should carefully check who was involved in the specific creative process and to what extent – and protect themselves with proper licence agreements.

Part I of this article explores the rights of authors in music generated by or with AI systems. This Part II focuses on further potential rights holders. Where musical works are created without a sufficient creative contribution by a human and no copyright arises, neighbouring rights can still exist in recordings, performances, broadcasts or databases involving such music.

Protection of phonogram producers

AI-generated music is protected and commercially exploitable if it is fixed on any sound storage medium. A phonogram producer within the meaning of Sec. 85 of the German Copyright Act (Urheberrechtsgesetz, “UrhG”) is the person who bears the economic and organisational responsibility for the recording of the sound material. The character of the recorded material is irrelevant. The protection covers recordings of musical works as well as any other natural or artificial sounds – including computer or AI-generated sounds.

Protection of artists

Artists (“performers”) have certain exclusive rights of use to their performances under Sec. 77/78 UrhG, including the rights of reproduction, distribution and making available on the Internet. Pursuant to Sec. 73 UrhG, a performer is anyone who “performs, sings, acts or otherwise presents a work or a form of expression of [traditional] folk art”. In the field of music, this typically includes musicians and singers. Other participants such as conductors and sound directors can also acquire these rights if they have sufficient artistic influence on the performance. Performers can only be natural persons – machines and computer programmes are excluded. This is because – similar to authors’ rights – the rights of performers protect a personal creative performance, namely the individual interpretation of a work.

The question of whether a human performance of AI-generated (parts of) music is protected under German copyright is more difficult to answer:

• The wording of Sec. 73 UrhG seems to exclude performances of non-human music. The provision requires the performance of a “work” (or a form of expression of traditional “folk art”). Pursuant to Sec. 2(2) UrhG, only an author’s “own intellectual creations” qualify as “work” under German copyright law. Autonomously AI-generated music does not meet this requirement since AI programmes cannot be authors within this meaning (see Part I of this article [link]).
• However, the legislator made it clear in the explanatory memorandum that the subject matter of a performance under Sec. 73 UrhG does not necessarily have to enjoy copyright protection. Rather, it also captures performances of works whose term of protection has expired or which cannot be subject to protection in Germany under international treaties.[1]
• According to German courts, the work performed must not reach the level of a personal creation under Sec. 2(2) UrhG, but merely be “suitable by its nature for copyright protection”.[2] It is likely that these courts would also grant protection to the performance of AI-generated music to the extent it is generally “suitable” to copyright protection. This applies to (parts of) music that would, hypothetically, qualify as “work” under Sec. 2(2) UrhG if it had been created by a natural person.

In the end, only the European Court of Justice (“CJEU”) can determine the scope of the performers’ protection. Most of the rights of performers are fully harmonised under Art. 2, 3(2) of Directive 2001/29/EC. Pending final clarification, the existing uncertainties should be taken into account when drafting licensing contracts.

Protection of organisers of a performance

If the performance of music is carried out as part of an event, the organising company acquires rights to reproduce the performance and communicate it to the public pursuant to Sec. 81 UrhG. The “organiser” is the company that bears the organisational and economic responsibility for the preparation and execution of the performance. To the extent that human performances of AI-generated music are protected under Sec. 73 et seq. UrhG (see above), the company organising the respective event can also acquire rights under Sec. 81 UrhG.

In contrast, the application of Sec. 81 UrhG to performances of AI systems must be rejected. The provision explicitly requires the performance of a “performing artist” which can only be a natural person (see above). While it is true that the organizer’s efforts are independent of the type of performance and the nature of the performer, the purpose of Sec. 81 UrhG requires a narrow scope: The provision intends to protect organizers who facilitate the personal interaction between an artist and the audience as it is valuable from a cultural and economic perspective.

As more and more cases will arise in the near future, it is to be expected that the German courts will have the opportunity to clarify questions regarding the organizing of AI music concerts. International and European law, on the other hand, does not provide for a specific right of the organiser of a performance.

Protection of broadcasters

If AI-generated music is broadcasted, the broadcasting organization acquires rights to the broadcast under Sec. 87 UrhG. As with the production of sound recordings, it is irrelevant whether the broadcasted content was created by a human or by an AI system. The broadcasters’ right is intended to protect their efforts in providing the program by investing considerable financial and technical resources. This does not depend on whether the broadcast involves a “work” within the meaning of Sec. 2(2) UrhG. Even the use of very short fragments of a broadcast may infringe upon the broadcasters’ rights, provided they are used in a recognisable form.

Protection of authors of a database

The protection of databases pursuant to Sec. 87a/87b UrhG can cover digital collections of AI music. Pursuant to Sec. 87a(1) UrhG, a database is a collection of “works, data or other independent elements”. Accordingly, the individual elements do not have to be “works” enjoying copyright protection – provided they are independent and systematically arranged. The author of a database obtains protection under Sec. 87a(1) UrhG if he makes a “substantial investment in terms of quality and/or quantity” for the obtaining, verification or presentation of the elements. Notably, however, investments that are only aimed at the prior generation of the elements later collected in a database, such as the generation of music by an AI system, are not to be taken into account.

Conclusion from a practical perspective

While music generated autonomously by AI programmes may often be free of any authors’ rights, performances and fixations of such music will have to be licensed by users. The rights of phonogram producers, broadcasting organisations and authors of databases exist irrespective of whether the respective music was created by humans or by an AI. As regards performances of AI-generated music and corresponding events, there is still a need for some clarification. Until the first rulings are available to provide guidance, artists, concert organisers and users should pay particular attention to the contractual protection of their interests.

From the perspective of users of music, it will be essential in all cases involving AI to ensure the greatest possible transparency regarding the stakeholders involved and the individual use of AI systems. Otherwise, it will be difficult to gain sufficient rights and agree on an appropriate remuneration. Supposedly “public domain” or “royalty-free” AI music should not be used in a commercial environment without contractual safeguards vis-à-vis the providers.

[1] BT-Drs. IV/270, p. 90.

[2] District Court Hamburg ZUM-RD 2010, 399, 409; District Court Cologne ZUM-RD 2010, 698, 701.