The last three quarters alone saw significant advancement in AI-generated music that exhibits remarkable complexity and emotional depth. Musicians are increasingly using deep learning algorithms to compose original pieces, leveraging vast datasets to generate compositions in various styles and genres. AI-powered tools and software have emerged, offering music executives the needed intelligent assistance in such areas as songwriting, production, and sound engineering.
These tools provide real-time feedback, suggest chord progressions, and even create accompanying harmonies in ways that represent the application of AI beyond the usual music recommendation systems and artiste discovery.
The disruptive nature of these technologies created uncharted terrains that sent the global music ecosystem into a scramble. Top labels like Universal have pushed for regulations around how AI is deployed in music amid concerns of potential infringements in synthesized vocals and systems trained on large volumes of copyrighted music. The likes of Sony hired an AI-specific senior executive to lead business efforts surrounding artificial intelligence.
The global industry is still divided against itself on AI music; structured markets like the US and Europe are still trying to work out guidelines governing AI usage in music, with the EU becoming the first major jurisdiction to pass a comprehensive regulation on the development, training, and deployment of artificial intelligence systems.
This has pitted the EU against tech companies. Ironically, the French streamer Deezer is using AI to identify and remove AI music from its platform while Spotify also opted to take down thousands of such songs from its site. AI’s disruptive influence on music production is reshaping the boundaries of human expression and altering the dynamics of the music industry. And Sub-Saharan Africa, with Nigeria as one of its key markets, is not left out.
Eclipse Nkasi, ex-executive at the Nigerian label Chocolate City, created a nine-track AI album that became the first of its kind in Afrobeats. But multiple lawyers say Nigeria is still far behind in regulating artificial intelligence and some specificities of the digital music ecosystem, even with the 2023 enactment of new copyright legislation.
At The Showcase Music Business Conference in Lagos, industry stakeholders of diverse specialities weighed in on policy and how Nigeria should be navigating the uncharted terrain of AI as it affects publishing and sync licensing. Among these executives are the intellectual property consultant Tope Salami, Tami Romeo who has worked with top artistes in Nigeria, and Janet Odunoye who is a publishing lead at the talent management company The Plug.
Stephen Kenechi: Availability and access to data that drives investment decisions is a major problem many business ecosystems beyond the creative sector struggle to surmount in Africa. You’ve had to touch on this topic as it concerns the music industry. Do you mind sharing some more insight, Tope?
Tope Salami: It’s not just in the Nigerian music sector, but the data gap is so obvious in Nigeria’s case because of its fast-growing nature. It affects the entire creative industry. It’s what everyone can observe. Most times, we play with data. When an artiste releases material, the number of streams and followership are all data points. Data is so important in the business of music because it gives room for investors and companies who want to offer you an advance in a deal to do their financial projections and determine what could be profited if the value is added to you.
None one person is to blame, but the entire system is culpable for this gap. The system has been designed in such a way that everyone believes you have to be in the top 100 to blow up as an artiste. I tell people that I’d prefer to be in the US top 300 or top 500 in the Latinos than for my song to enjoy the number 1 spot in Nigeria for a month. I’ve made use of this example because royalty payment schemes vary in their nature, and that’s because there’s a gap in data generation, collection, and consumption in Africa and particularly in Nigeria. It’s not that bad in the advanced world. That’s why you’d see people who release only one song in their lifetime doing better than someone in Africa who has been releasing two songs per month. The data over there is accurate and relatively better and it gives room for reliable projections. This doesn’t affect the mechanical rights segment alone. It also hurts our music publishing.
Imagine that I’ve registered my song to earn publishing revenue. What data have I been supplied or did I supply? You can’t give what you don’t have. Technology-driven systems will definitely process what you’ve supplied them in conjunction with AI schemes. Results will be generated in the form of data based on what was supplied. Not to be alarmist, but we’re getting there. New companies are coming on board every day. Practitioners like us are shedding more light. In the end, our veterans who paved the way for the new generation won’t find themselves in situations where they are out of cash. With data, you should know how much you will make in a given year. I know I can’t fall below a particular equilibrium in a given quarter because certain things have been done right. That’s data at work.
Even your name as an artiste is data. Imagine searching “Davido” on Spotify but you erroneously write “Da Vido”. You will probably see someone in Spain with the name. Assuming I don’t understand how these work, I might download the person’s album thinking I’d downloaded Davido’s work. I later realised I’d downloaded the wrong thing, adding to the revenue and fanbase of the artiste involved, to the potential loss of the real artiste intended.
Data can solve these challenges. We need practitioners who appreciate data and would not break professional rules. Data goes deep. Imagine that I get to the airport in the US and they’re doing my tax return. That’s because they have accurate data. Lagos state is trying to tax our creatives. Do they have accurate data? No! If I’m a creative. How do you know how much I earn based on my work as a creative or businessman? This causes unwarranted strife.
Stephen Kenechi: So by data, you mean holistic streaming stats that can be exploited for predictive analysis, publishing revenue data, industry data that can serve as a framework for taxation? Music metadata? What else?
Tope Salami: There’s performance data as well, which many people don’t pay attention to. If I can sell out an entire 70,000-capacity concert arena for three days, that’s 210,000. It shows that I need a bigger space. If I get a conservative 150,000-capacity arena, I should be able to sell that out in one day. These analyses are quite simple. Those who run such concert spaces as the O2 pay certain fees and taxes to the government based on data. Concert promoters can make informed decisions, like, “Maybe we’ll use 900,000-capacity Wembley next.” It’s all data.
Stephen Kenechi: Speaking of tech-driven systems, generative AI continues to gain new ground in music. What implications do you see in this as it affects rights assignment, IP management, and potential infringements?
Tope Salami: AI has always been there since the inception of technology, but people probably didn’t understand the concept. Our mobile phones and applications have AI functionalities. Yes, AI has been doing a lot in the music space. But whatever it is we’ve achieved with it, there’s a need to regulate those or risk our creatives losing value. There are many implications, the topmost of which is a potential infringement on rights. There’s a natural way of expressing pain which is conveyed in songwriting and musical releases. You get to act with discretion. AI does this, but subject to the coding written for it. Practitioners increasingly relying on AI are likely to lose creativity. AI will definitely be abused. A case in point is stream farming where AI bots are used to push songs into topping charts. Even a terrible song would raise brows and attract revenue this way, making entries with organic streams lose their chart slots too early. Artistes displace each other easily on the charts with this, showing how weak the system is.
There must be regulations for artificial intelligence, just as EU nations are trying to create systems for this. There will be loads of potential infringement, hence stakeholders must come together to see that things are done right.
Janet Odunoye: I think AI-assisted music production can be a good thing, but the implication is foremost in how it complicates copyright ownership and authorship. AI gets to generate content without human error, creating confusion as to who is the author of an AI-generated musical work. That’s why we need to adjust our legislation to account for the copyright of such works. Who owns it? Is it the platform with which the music was generated or the human effort that prompted it? Also, in terms of royalties, who earns what for what right? The traditional norms of distributing royalties will definitely have to evolve to accommodate AI-generated content. How about sampling and interpolation; using someone’s pre-existing work to create your own with AI? There is a serious implication here because generative AI is essentially trained on pre-existing music or just copies from old works. Should this amount to an infringement? Who would be liable for such infringements? Is it the human component or the platform used?
Stephen Kenechi: Synthesized vocals are an uncharted terrain in global music, including much more structured industries like the US. How should the local music scene in Nigeria be responding legally and business-wise, Tope?
Tope Salami: As you said, these things are already happening in top countries. Whether we like it or not, the big players have entered our market. They have what we don’t have. They’re coming with the money and the back-end technologies. Whether or not we preach regulations, AI will happen. The only way we wouldn’t be caught unaware is if we realise early enough that these technologies are unfolding on a global scale and will hit us sooner or later.
Tami Romeo: One thing we can consider in regulating the use of AI in music production is creating guidelines that speak to transparency; disclosing the extent to which it was used in production or post-production. It should be clearly stated if a project had 50% AI input. This could make people become comfortable enough to consume or make that kind of music. We could also be creating awareness so that people will know that AI isn’t here to stifle creativity but to improve what has already been created. Creatives face mental blocks and could be turned off from time to time. AI could help point them in the right direction. It’s not something we should be afraid of embracing.
We should start having policy talks on this with producers, music business executives, and other stakeholders who are involved in the creative process so we can understand where we’re coming from, where we are, and where we’re headed. We know that the Copyright Act was just signed into law. Now is the time to start speaking on these issues so that, in the next two to three years, there’ll be an amendment. It’s a collaborative effort between humans and AI.
Stephen Kenechi: Tami, in what specific ways have creatives in Nigerian music been deploying AI?
Tami Romeo: I have a very great example and he’s my friend and a veteran artiste called Eclipse Nkasi. Recently, he released an AI-generated Afrobeats album called ‘Child of Destiny’. This album was created in three days. It has the infusion of his local dialect Igbo in it. The music, lyric, and beats were all written by AI. When I listened to the album, I was so impressed. He’s the first Nigerian artiste that I know of to do that. It made me see the other side of embracing AI in music. It made me see that there needs to be a significant amount of human input as much as AI in the generation of that music. He had the idea to do this and had to find the exact tools he needed, input the right commands, and generated what he wants. It’s a step in the right direction. If people begin to look at using AI in such a positive way, we will see it as a normal thing in the next two to three years. AI has always been there. Social media algorithms work with AI. For this feat to have been achieved by a Nigerian means we’re interested in getting things done the right way. I believe Nigerians are someday going to be the front-runners of AI in music production.
Stephen Kenechi: Janet, how much of a dilemma do you think AI-powered streaming fraud and farms pose in the Nigerian market? And in your opinion, just how much of the streaming stats we see in Nigeria are organic?
Janet Odunoye: I don’t think they have to exist in Nigeria to become a concern. The world is now a global village. You can do anything from anywhere. You can pay for things to be done in Nigeria from other locations in the world. These are unethical ways of hijacking charts and amassing royalties, denying artistes who are actually putting in the work a real shot at revenue and recognition. Tackling these is firstly a burden for the streaming platforms who have to put extra security on how creatives sign up on or use their platforms. It could be an authentication or ID system. Streamers should work together on it. Collective management organisations, music labels, publishers, and other stakeholders after organic artistes should weigh in on this. This is an aspect of music that could be covered by laws, but even if there’s a local law against this, this law would probably not cover jurisdictions outside Nigeria. It’s a topic that we can keep enlightening artistes on. It says a lot about brand integrity. You won’t top the charts forever.
Stephen Kenechi: You vaguely spoke of dysfunctional performance rights societies, specifically mentioning the peculiar situation of MCSN in Nigeria. What is the status quo and what key reforms are needed to turn it around?
Janet Odunoye: MCSN is the performance rights society for musical works in Nigeria, but it’s not doing the job enough, even after the battle with COSON to be legally recognised as the sole collector. If Nigerian artistes want to get their royalties, they still have to register with international collective management organisations like PRS, BMI, PPL, and ASCAP. If things worked as they should, foreign collectors are meant to be remitting royalties to MCSN who will then transmit this revenue collected across the world to Nigerian artistes, songwriters, and composers.
It should be a collective effort between MCSN and collective management organisations abroad to get the royalties due to Nigerian artistes. This shouldn’t even be a conversation. They should be actively doing it. It shouldn’t be the case where local artistes here are having to register with three to five performance rights or collective management brands to get royalties. There should be a proper structure as to how our artistes get these royalties from MCSN.
Tami Romeo: In the age of AI, music rights administration has to evolve. Our existing licensing models may need to change and adapt to accommodate AI-generated music. Licensing organisations and platforms can work towards developing specialised kinds of licences that account for the unique characteristics of AI-generated music content. They can stipulate associated royalties and rights, just so we can ensure there is fair compensation for all parties involved, both human creators and AI developers. We can also consider revamping the rights administration on the collective management organisation. CMOs most likely will need to update their systems and processes to handle AI-related rights. They can include establishing systems that ID AI work in production, vocals, and general masters of the song. It’s an ongoing evolution. A monitoring system that adapts as tech evolves will come in handy because it will be important to monitor the landscape of AI-generated music in the next five to ten years. This will help us assess the strength of the regulations we’re adopting, structure policy frameworks, and get more out of what we do.
There has to be a close collaboration between legal experts, music producers, artistes, and all industry stakeholders.
Stephen Kenechi: So where is Nigeria in this global scramble to protect artistes and songwriters against AI infringement?
Tami Romeo: Nigeria is not yet at the forefront. The Copyrights Bill of 2022 was only assented to in March 2023, repealing the Copyrights Act of 2004. The conversation to repeal the act started long ago and we’re just getting to it. We’re still off when it comes to policymaking. But I want to acknowledge that the conversations we’re beginning to have in the music space are the kind that can turn things around in the next three to four years. Stakeholders at every level are researching and it’s a great sign. It pushes us to look into things like this and join the discourse so we don’t get left behind and have to catch up in the next ten to 15 years. I see a shift happening in two to three years when this discourse becomes more mainstream in our music generation, sound engineering, and production.
The first edition of The Showcase music conference is a starting point that made me see that people have interests, thoughts, and opinions on this topic. If we continue to have this conversation, we may be able to reshape the laws.
Copyright 2023 TheCable. All rights reserved. This material, and other digital content on this website, may not be reproduced, published, broadcast, rewritten or redistributed in whole or in part without prior express written permission from TheCable.
Follow us on twitter @Thecablestyle