The impact of AI on the music industry part 2: Q&A with researcher Rachael Drury | ISM

Can you tell us a bit about your research and how you came to work with the ISM on this survey?

My research looks at the commercialisation of generative AI music (machines that use large corpora of music as data and algorithms to generate new music) and platforms which can be used by anyone to generate music faster and more cheaply than a human. More specifically, I am looking at the challenges that generative AI brings to the UK music industry’s copyright framework by considering the authorship and ownership of AI generated music and the potential infringement of copyright protected music used in the training of AI: both challenges have the potential to impact the value of human-created music.

Understandably, there is concern that AI may end up replacing the work that composers and songwriters rely on for a steady income. The ISM represents and promotes the interests of music creators, while providing a link between musicians, researchers, government and policy makers. It is hoped that my research will support the ISM when they look to action change through discussions of law, policy and regulation with government, but also that it will provide information and advice to music creators moving forward. The ISM has been incredibly supportive of my research and continue to promote my research survey which aims to better understand music creators’ views and opinions on AI in the music industry.

Many of the questions about Intellectual Property and Copyright Law intersect with similar issues raised by sampling and streaming. Could you tell us more about this?

Yes, both sampling and streaming raised questions around copyright infringement and led to legislative changes that increased protections for music creators.

During the early days of sampling, copyright law didn’t specifically address the use of samples and so sampling was practiced freely without any permission or payment to the original copyright owners, just as music (in some cases copyright protected music) has, to date, been used freely as training data for AI. A number of lawsuits ultimately led to the industry-wide standard that licences are required for the use of samples, and it is likely that the future use of music as training data will also be decided in court.

The advent of streaming addressed the problem of infringement in the form of digital online piracy, but many music creators have been adversely affected by streaming through oversaturation of the market and poor standards of remuneration. When we look to find ways for the music industry to adopt AI, we must make sure that music creators are protected and fairly compensated if their work is used to train AI systems.

Anybody interested in how experiences with sampling and streaming could help us to move forward with AI can read my blog post on this very subject!

The survey aims to investigate music creators’ views and feelings towards AI – why do you think this is important?

It is music creators who will be using AI tools in their creative practice, and it is music creators who will feel the impact of AI music platforms that could replace them in the workplace. This survey will help us to understand if there is a place for AI within creative practice and therefore an opportunity for music creators to work together with AI companies to create products that are useful. Additionally, copyright is a framework that provides protection and financial reward for creative output. Music creators deserve to voice their opinions and concerns over what any changes to copyright law could mean for their income and future careers. This survey offers music creators the opportunity to have their say. With the findings, I hope to share the voices of music creators with policy makers while providing a balanced and fair assessment of how the UK should proceed with regards to changes to copyright law and the regulation of AI.

My research survey can be accessed here.

What are the most important/interesting points you’ve taken away from your research into music and AI so far?

There are two general points that really stand out to me. The first is that interest in AI music has grown exponentially since I started researching this topic back in 2020, just as AI music generation platforms became viable commercially. The second is that from AI companies to policy makers around the world, nobody really seems to know the rules or the answers; creative AI remains a complex and controversial topic that is going to take time to resolve.

It has been an interesting journey as a researcher to speak with many music creators. The initial reaction towards AI was one of disdain and fear; there has always been a perspective that creativity is a human endeavour that AI has no place in simulating. However, as the hype begins to settle, more and more music creators are asking how they can use AI to their advantage. Increasingly, there seems to be the feeling that you either get onboard or you get left behind.

As an expert in AI and music, what advice would you give music creators to best equip themselves for the change that is afoot?

Of course music creators should be wary of AI. Inevitably, parts of the music industry will be severely disrupted by this technology, and music creators should strive to protect themselves and their work. Make sure your voice can be heard. Engage with organisations like the ISM, write to your local MP and get involved in government inquiries.

However, I think it’s important to understand that AI isn’t going away. Be willing to educate yourself about the AI tools already available to you; in some cases AI may enhance your creativity and your workflow. The onus shouldn’t be solely on music creators however, AI companies should work with music creators when building tools and provide the information and education needed to use the tools effectively so that creators can understand their value.

AI is a technology that should be embraced when it proves to be useful. I think there are sounds and genres of music out there yet to be discovered through the use and misuse of AI. Find these tools, learn about them, play around with them, misuse them, break them, see what you can do with them! Do what you do best, be creative!

What benefits (if any) do you think AI might present for musicians and people working in the music industry?

Many tools can help streamline the creative process, from DAW plugins that can find you samples, create you a virtual drummer or vocalist, and generate melodies or chord progressions, to AI tools that can catalogue your music and organise your files. AI mixing and mastering software is also available, and some of the most intuitive instrument modelling powered by AI can give you a better idea of what your finished product will sound like when performed by humans. There are hundreds of AI-powered applications out there, but every music creator will have different needs and some applications will be better than others. Decide what you need from AI and see if there is a tool out there that can help.

What do you think the role of bodies like the ISM is in this issue?

Industry organisations like the ISM are vital in giving a voice to those who create the product on which the entire music industry is built. Without music creators there is no music industry, and without human music there is no AI music. It is vital that music creators are listened to, looked after and spoken for in places where they can’t be heard. Organisations like the ISM have a direct line to policy makers and regulators in government and they can speak on your behalf; they will fight your corner.  

Are there any further resources you’d recommend to musicians to learn more about AI and how it may affect their work?

There are hundreds of articles online that discuss this very issue, but many are biased either towards the music industry or towards AI and some are simply misinformed about the reality of the technology’s impact. The positive is that there is far more research happening now that will begin to answer some of the questions that music creators have. Water & Music has a regularly updated archive of articles on AI and music as well as other areas of music and technology. I’d also recommend engaging with resources by music creators who are actively engaged with AI. Holly Herndon, for example, builds her own AI for music creation and hosts a fantastic podcast with her partner Mat Dryhurst called Interdependence in which they discuss the technical, ethical and legal challenges of AI and music.

Read Part 1 of this series here

NMC's Discover platform is created in partnership with  ISM Trust.