Art-ificial for the Sake of Art-ificial: How is AI Changing Copyright Law?
/Wavo Blog: This series shares information that’s most relevant to music marketers around the globe.
Context
Declaring copyright ownership is becoming more complicated. Development in Artificial Intelligence has led to essays, paintings, poems, music, and more, being generated by AI machines. These machines are not perfect and there is plenty of room for improvement. However, these AI-generated songs are improving every day. But what is the implication for the music industry?
On the one hand, this could allow musicians to create with AI and perhaps make even better music. On the other hand, it also poses a threat to ‘organic’ songwriters, who create their music without AI. The worry is that AI music generators become so good that there is no room left for human songwriters.
AI is offering technology that efficiently “helps” songwriters to make the song they want. Songwriters still have creative input but AI makes the process easier - offering them programmes and virtual instruments to improve songwriters’ own ideas. AI can even analyze specific recordings to regenerate a particular melody.
It’s not just the composition either, a songwriter may use AI song generators for lyrics too. The latest example, ChatGPT, is not only able to write poetry, but even write in a similar technique to specific poets. Subsequently, a songwriter who just needs the lyrics to finish their song, can ask ChatGPT to write x number of verses on y topic, in the style of Leonard Cohen.
It may seem as though the (human) songwriter is no longer needed with the potential to make hit songs easily with AI. However, integrating AI into music production is not straight forward.
Copyright
Copyright is the foundation of royalties and ownership in the music industry; and Copyright Law governs how artists can declare Copyright ownership of songs. The Copyright process is already complicated, but it is getting even more difficult to navigate as AI generators wade in on the songwriting process. A study carried out by Wolters Kluwer on how AI music generators apply in EU law begins to answer this question:
Does AI output qualifies as a work from the perspective of EU law?
The four interrelated criteria that answer whether an AI output qualifies as a copyright-protected “work” are as follows:
A “production in the literary, scientific or artistic domain”;
The product of human intellectual effort;
The result of creative choices;
The choices “expressed” in the output
The analysis is that many of the AI music outputs examined will likely pass steps (1), (2) and (4) unscathed. The crux of the test is therefore in step 3, which encapsulates the essence of the originality standard under EU copyright law.
From this perspective then, where an output does not qualify as original in the sense that it reflects the author’s free and creative choices, that output is - from the perspective of copyright - in the public domain.
One of the conclusions Wolters Kluwer deduce is that it is useful to distinguish between two scenarios: (i) when AI developers and users are the same person(s); (ii) when AI music creation systems are offered to users “as a service”. The second scenario is more complex for establishing copyright ownership of the final product, a song. On the one hand, how much of the song is written by AI; and, on the other hand, the creative choices made by musicians to make the song. It seems as though this second scenario can potentially fulfil condition 3 as well as criteria 1,2 and 4.
However, it comes down to the extent of the AI’s influence in the creative process and how much work the songwriter has contributed to produce the final work - something that is likely to be decided on a case-by-case bases within this current legal framework.
AI technology is proving it can already write lyrics, composition and even generate an artificial human voice. We aren’t far away from AI artists taking over streaming platforms and charts with very limited human creative input. If human songwriters become more irrelevant, the focus will turn to a new question:
Is the developer(s) of the AI entitled to the copyright of a song that was entirely generated by the machine; or, is the AI itself the author of the music?
According to the CEO of SoundExchange, Michael Huppe, we are already at this problem with the non-human composer Artificial Intelligence Visual Artist (AIVA) “composing classical music by analysing the patterns in 30,000 scores.” With AI machines, such as AIVA, producing excellent songs there is perhaps an argument for these machines to claim Copyright ownership instead of the people who built the machines.
Key Takeaway
Although the discussion will centre around whether the AI or the developer deserves the Copyright, these cases will also reopen the recurring debate over what credit is due to those scores that ‘inspired’ the AI? Copyright in music is dominated by plagiarism disputes with stars such as Ed Sheeran, Radiohead, and Led Zeppelin getting bogged down in years of legal battles. The accused in these instances often defend themselves by claiming ignorance, coincidence, or subconscious inspiration. But what of when we know for certain that an AI machine has analze the past works of songwriters to invent music? Will the composers/owners of the songs that the AI analyses to make a new song demand song credits and royalties for their part?
Whatever the outcome, Copyright law is being challenged and governments around the world will likely need to legislate to accommodate non-human songwriters/artists. One day soon AI artists might dominate the charts and these legal definitions could determine who owns the future of music.
Written by Rufus Darwall
Join our Newsletter
Join thousands of music marketers who read our weekly newsletter to stay a step ahead on music industry trends, marketing, and analytics.