Why Generative AI is Bad for Music Production

Generative AI is Changing the Game for Music Production:

When I was a young kid, my piano teacher always said that practicing the basics would make performing the “hard stuff easier.” Little did my teacher know that more than two decades later, the average musician could complete the “hard stuff” without touching a single instrument using Generative Artificial Intelligence (AI). The recent adoption of AI technology for music production by companies like Apple, Adobe, IBM, and AWS has shifted the music production market towards faster, more precise, and accessible digital workstations . Furthermore, the growth away from live music due to the global pandemic has accelerated bedroom producers' use of Generative AI-based music production platforms. However, the widespread adoption of these platforms has not come without controversy since many of their functions automate the skills that traditional studio engineers have perfected for decades. In this article, we’ll explore why Generative AI could forever alter aspects of the music industry that have been left unaltered for decades.

Creates a “Wild West” for Music Sampling:

While the Music Industry has long battled the proliferation of illegal sampling, using Generative AI to produce music opens the door for promoting a “wild west” for music sampling that could have unintended consequences for proliferating copyright infringement and illegal downloads. When Vanilla Ice’s famous single “Ice Ice Baby” was released in 1990, Queen and David Bowie clarified that simply adding notes to a common baseline does not make that sound a unique work. However, music production tools that leverage AI often replicate such a process using reinforcement learning with similar harmonic and melodic patterns to produce “original pieces of music.” As music production platforms adopt AI, the music industry must consider drawing other lines to distinguish between “original” and “sampled” works.

Makes Music Licensing Obsolete:

Metadata has played a vital part in enabling music producers to get fairly compensated for their work over the last two decades. However, Generative AI platforms often acquire large catalogs of “royalty-free” samples that use anonymized data from blanket licensing agreements that often prohibit producers from getting compensated fairly for their work by receiving publishers. As a result, users are provided unfettered access to infinite amounts of music samples without properly paying the producers behind them. As Generative AI plays a more significant part in music production, the industry will need to determine who has the right to use this material and how the creators of such material will be compensated fairly for its use.

Poses Privacy Concerns for Music Producers:

Producers that want to “make it big” must advertise their work. However, many AI tools pose privacy concerns to bedroom producers by creating a veil of secrecy around how their data is stored and exchanged with third parties. Most AI tools focus about 64% of their processing power on automating data collection, which means AI music production tools consume significant amounts of personal data to produce original musical content. Since AI music platforms sit as interfaces on top of large servers, many producers may not know how much data is collected on the thoughts, feelings, and emotions behind their music. Therefore, music producers must exercise caution when using these platforms to ensure that the AI tools don’t collect sensory data that must remain private.

Opens the Door for Intellectual Property Theft:

AI music production platforms don’t always consider the risks of digital Intellectual Property (IP) theft when making their software available using open-source protocols. AI music production platforms that use open-source protocols face significant exposure to Distributed Denial of Service (DDoS) attacks that leave users’ data vulnerable and open the door for digital intellectual property theft. Unlike decades ago, where IP theft required unauthorized access to the physical medium, a simple DDoS attack on an AI music production platform could not only hand over sensitive user data but also enable hackers to obtain early access to the users’ content with a simple click of the mouse. While there are moral advantages to “democratizing” the use of Generative AI music production platforms, there are also important cybersecurity challenges.

Generative AI for Music Production is Risky Business:

While Generative AI provides a new way of creating and advertising musical content, there are also significant concerns around the need for more privacy and IP protections that must be addressed to ensure that such tools provide the maximum benefits to music producers and recording artists. Without a legal and ethical shift in how the music industry interprets sampling, Generative AI will continue to open the door for unregulated access to music samples. Furthermore, using Generative AI in music production raises significant privacy concerns about how users’ data is stored and shared with integrated applications. Finally, the use of Generative AI to produce music highlights the need for more excellent encryption and cyber protection solutions for open-source technologies to ensure that a producer’s work is not illegally obtained and distributed. The role of Generative AI in the music industry is only expected to grow in the coming years; therefore, as music producers introduce the latest content to the marketplace, they should keep in mind the risky impacts AI can have on creators, performers, and consumers alike.

Sources: Forbes, The Verge, DJMag, Atlantic Economic Journal, BBC, TechCrunch, Splice, Music Business Worldwide, Gartner

About the Author: Jonathan Sherman is a Senior Business Analyst in the Music Industry, where he builds rights management and publishing analytics tools. He co-founded one of the Big 4's Music Industry Consulting Practices and spent over a decade in the music industry as a producer, disc jockey, label executive, and analyst. Jonathan has spent a decade in management consulting in business strategy, digital transformation, business intelligence, and intellectual property management. He holds a BA in International Economics from American University and is currently researching the intersection of media and technology as a graduate student at Georgetown University.

The ideas, views, and opinions expressed in this article represent the author's personal views and do not reflect those of the author's current or previous employers.