Analysts predict a tenfold increase in revenue for the artificial intelligence music creation market over the next 10 years. The subject is becoming increasingly controversial, related to issues such as algorithmic learning from already composed works or using vocalists’ voices to create covers sung by AI. At the same time, over one-third of musicians admit to having used artificial intelligence in their work. However, artists expect clearly defined regulatory frameworks for such tools, primarily concerning permissions for the use of intellectual property and clear labeling of AI-composed music.
“We observe the presence of artificial intelligence in all areas of our professional lives, and the music market is no exception. Artificial intelligence enters on two levels. On one hand, in terms of music production, and on the other hand, as support for various supporting tools: whether in the music publishing process, music promotion, or simply conducting business. We observe many spectacular cases of using artificial intelligence in music, not necessarily according to the expectations of the authors and creators of musical works,” said Anna Ceynowa, Communications Director at ZPAV, in an interview for Newseria Innovations.
On one side, algorithms are used by streaming platforms to recommend tracks that listeners might like. However, AI often inappropriately draws from the artistic work of musicians. One of the most striking cases recently was the hit “Heart on My Sleeve,” published on streaming platforms by user Ghostwriter977. Based on the vocal, listeners attributed the song to Canadian rapper Drake or a Canadian singer using the pseudonym The Weeknd. It turned out, however, that the popular song was generated by artificial intelligence, and the vocal profiles of real artists were used by the algorithm to generate the voice.
There is also much controversy regarding the AI-generated songs or performances that mark the voice of deceased vocalists, such as a cover of “Bohemian Rhapsody” from Queen’s repertoire, sung using the vocal profile of Chester Bennington from Linkin Park, who died several years ago.
“As we hear various kinds of new song remakes, which are performed using the voice of known, liked, popular artists, but that voice is generated by artificial intelligence, these are curiosities that definitely arouse interest. But I think there is a lot of work to be done in terms of education and building awareness, also among the recipients, regarding the respect for the rights of creators, people who create music and live from the work of creating music. It is very important that we as recipients are aware of this and can distinguish. And if we are not able to hear the difference with our own ears, then we should be informed whether the musical content we hear is the work of a human, or AI-supported,” believes Anna Ceynowa.
The study “Engaging with Music 2023” conducted by IFPI, the international organization representing the recording industry, shows that music listeners are already highly aware of the existence of artificial intelligence, many of them use it and are interested in its capabilities. However, nearly eight out of ten believe that human creativity remains essential for creating music. The vast majority of respondents agree that AI should not use music or an artist’s voice without their consent, and that artificial intelligence should not be used to clone artists or impersonate them without permission. They also recognize the need for transparency and the establishment of rules for AI systems in the music sector.
“Certainly, a debate and understanding are needed. As a rule, however, there is a belief that artificial intelligence will not replace the creator, will not replace the artist, will not replace talent. Certain processes can be supported, facilitated, or copied thanks to artificial intelligence, but it will not copy uniqueness, talent. Human imagination is so extensive and unlimited that these concerns in the industry are not dominant,” reassures the ZPAV expert.
According to a study conducted by the Goldmedia Institute on behalf of GEMA and SACEM, 35 percent of music creators in France and Germany have already used some type of AI tool in their work. Among creators under the age of 35, this percentage rises to 51 percent. Of all respondents, 19 percent stated that they do not consider using artificial intelligence at all.
At the same time, 95 percent of respondents expect all cases of using copyrighted works in training algorithms to be disclosed. 90 percent also want creators of artificial intelligence to ask for permission before using copyrighted works in training.
“Regulations are needed, but they must be preceded by a long and deep debate, so that we do not over-regulate the market. Protection of creators is absolutely necessary. Such directional guidelines are already included in the AI Act, adopted by the European Commission. Above all, access to information about the musical content used by machines learning from existing works to generate new content is also necessary. The need for labels or sharing information on which copyrighted works and content various AI-based tools use is one of the essential elements of such a system that should definitely be regulated,” believes Anna Ceynowa.
According to a study conducted by the Gold media Institute, revenue for music creators could drop by as much as 27 percent by 2028 due to the development of artificial intelligence. In contrast, the market for artificial intelligence in music is expected to grow tenfold from 2023 to 2028, reaching $3 billion.
