AI has become a ubiquitous topic, infiltrating every facet of our lives, including the music industry. While there are debates in favor and against AI in all industries, it feels like arguments against AI are felt strongest in the music and creative industries.
For the music industry, AI offers the prospect of streamlining operations and improving efficiency. However, the future of AI in music creation, copyright, and consumption is widely debated. That’s why in an era where the industry diligently worked to ensure fair compensation for rightsholders, the introduction of AI raises the specter of a return to the challenges posed by illegal downloads over a decade ago.
AI and copyright infringement
One of the most pressing concerns surrounding AI in the music industry is the potential for copyright infringement. AI algorithms can generate music that closely resembles existing works, raising questions about ownership and originality. Rightsholders and artists rightfully fear that their creations could be exploited without proper compensation or attribution. This issue has already led to legal battles and calls for stricter regulations to protect intellectual property rights. In fact, earlier this year Tennessee became the first US state to pass legislation to protect musicians from unauthorized AI impersonation, while the RIAA and the major record labels filed a lawsuit against generative AI audio startups.
While AI is highly contested on the “fake” impersonation and copyright infringement side, it also has the potential to be a powerful tool for creativity and innovation. AI-powered software can assist musicians in composing, arranging, and producing music, it even can help identify emerging trends and preferences, enabling artists to tailor their music to specific audiences.
AI is not new
On the consumer side, AI has been in action long before the words “generative AI”, “OpenAI” and “ChatGPT” made headlines. AI and data science has been in use for a long time – it’s just not the futuristic, dystopian type that we see in scaremongering articles. Algorithms – or “recommendation engines” – that recommend new music, and movies, display relevant ads on social media, or an Uber nearby – these are all powered by algorithms that are based on data science. Through AI, DSP platforms can improve the user experience and match their listening habits with new, undiscovered music. DSPs analyse audio features, metadata, and user interactions to recommend songs or artists matching a particular mood, tempo, or style and offer personalized recommendations.
Arguably, one area where AI could do the most good for the music industry is one that is hardly noticeable to creators and audiences. That area is the industry’s supply chain. Why? First, let me explain that I’m not talking about the supply chain in the sense of the distribution of music from a record label to physical and digital stores, and DSPs (although it is related). We’re talking of the supply chain in terms of the data and metadata that is associated with music works and sound recordings. The data associated with that travels in various directions (e.g. radio, DSPs, physical sales, sync) and the challenge is to match the use of these works with the work itself and then with its rightsholder. If these matches don’t take place correctly and efficiently, the royalties to artists and rightsholders will be either delayed or never arrive.
Behind every track, there is data – metadata – that can identify a track and match it to its rightsholders. It is not just the name of the song, artist and album. It goes into far much more detail. This includes songwriters, producers, publisher(s), record label(s), copyright holder(s) and of paramount importance: essential identifiers like ISRCs, ISWCs, and IPI numbers. And because there is no international standard (yet) a lot of this information is missing or gets entered incorrectly as it moves between various databases. Each database can have a different format of information which causes the various works out there in the supply chain to not communicate with each other.
AI and metadata
The good thing about AI is that it can handle vast amounts of data. Record labels, publishers, collecting societies and other music administrators handle hundreds, if not thousands, of lines of data on a regular basis. AI can make metadata management much easier and remove the human error that can be so easily introduced in our daily use of spreadsheets. AI algorithms can improve the accuracy and consistency of metadata, reducing errors and streamlining the royalty distribution process. A higher metadata matching percentage leads to a more equitable distribution of revenue among rightsholders.
It can go even further – machine learning can also be harnessed to identify patterns in data, improve metadata matching, de-duplicate, and even detect anomalies that might indicate lost royalties or fraud. All of this can have an impact on rightsholder royalties, the amount artists get paid and the speed with which they receive their money. On the business side, it can also help reduce operational costs and hours lost tracking data in spreadsheets, helping them focus on other important aspects of their jobs.
AI isn’t the only solution – we need industry collaboration
AI could be the closest to a solution to the metadata problem. But to get rid of the problem entirely, the industry needs to work together. Similarly, the industry needs to come together when it comes to AI. The key to harnessing all of the aforementioned AI benefits while mitigating its risks lies in finding a delicate balance between fostering innovation and safeguarding artists’ rights. This requires collaboration between industry stakeholders and policymakers.
To harness the potential of AI while mitigating its risks, all stakeholders in the music industry must unite to establish ethical guidelines and standards. This necessitates defining AI-generated content precisely, ensuring transparency in AI usage, and safeguarding the rights of artists and creators. Tech companies developing AI-powered systems bear a crucial responsibility to adhere to these guidelines. The industry cannot afford to wait for legal frameworks to evolve, risking a repeat of the Napster debacle.
Jacob Varghese is a technology leader and founder of Noctil, a music and audiovisual metadata management platform. With over 24 years of experience leading complex technology projects, Jacob started Noctil to fix the inefficient systems and processes that cause delayed or lost royalty revenue in the music and audiovisual industries. Jacob strives to ensure rightsholders are paid for their work, and make technological advancements and research accessible and affordable to the industry. Working with Music Licensing Companies (MLC) and providing technology leadership to build global repertoire metadata has given him an in-depth understanding of the music and audiovisual rights and licensing process. Before Noctil, Jacob has worked in the telecom, media, and entertainment industries with Fortune 500 companies. Jacob is passionate about using technology to solve complex business problems and drive sustainable growth. When he is not immersed in the latest tech trends, he can be found mentoring aspiring entrepreneurs or exploring the great outdoors. He has a graduate degree in Mathematics and Computer Application.
