MBW Views is a series of exclusive op/eds from eminent music industry people… with something to say. The following op/ed comes from Nick Breen, a Partner in the Entertainment and Media Industry Group at Global law firm Reed Smith.
Artificial Intelligence is beginning to play a significant role in the music industry, marking a considerable shift in how we create and interact with music.
The recent creation of a “final Beatles song” by an AI model shows the capacity of AI to generate new material, even mimicking the distinct styles of heritage acts who are no longer with us.
This level of advancement serves as a testament to the growing power and sophistication of AI technology in the realm of music.
However, AI’s transformative potential brings an array of challenges to the industry. The principal issue is finding a balance that promotes technological innovation while protecting the rights and contributions of human artists.
Copyright: Navigating the Landscape
The copyright ramifications of AI in music are as complex as they are impactful. Current copyright laws provide a framework, but they face limitations when applied to the evolving field of AI-generated music.
AI systems require vast amounts of data to develop, often consisting of copyrighted works, creating a maze of legal questions.
Can an AI lawfully ‘train’ on copyright protected music?
The answer, echoing the favoured response of lawyers everywhere, is: ‘it depends’. To train an AI model, a copy of the training set needs to be made. Where the training set contains content protected by copyright then the default position is that authorisation is required from the copyright owner, unless a copyright exception applies.
Copyright is territorial in nature and different regions have adopted varying stances on the issue. For example, countries like Japan, Singapore, South Korea, Taiwan and Israel have introduced exceptions to their copyright laws that favour AI developers.
In contrast, the EU has adopted a similar approach, albeit with a significant caveat allowing copyright owners to opt-out. The UK initially proposed such an exception but later withdrew following substantial opposition from the creative industries. Across the Atlantic, there appears to be US case law that would support the position that copies made to train an AI is covered by the doctrine of ‘fair use’, though such position is already starting to be challenged.
Is AI-generated music protectable?
International copyright law, as per the Berne Convention, necessitates the involvement of a human author for a work to be protectable. This stance aligns with the U.S. Copyright Office’s recent guidance. In short, a work predominantly generated by an AI model likely doesn’t qualify for copyright protection. However, if a human author uses AI merely as a ‘tool’ to create a work, that author might still enjoy copyright protection. The pivotal question then becomes: how much human involvement is required? This query will be a subject of intense debate in courtrooms across jurisdictions for many years to come.
As you can see, the responses to these questions are far from straightforward. The position is constantly evolving and further complicated by various political and jurisdictional factors, where the friction between the interests of the technology industry and the interests of the creative industries takes centre stage.
Various stakeholders, including artists, record labels, digital service providers, and AI technology companies, are adjusting to this new landscape. Each group is formulating their position, updating their practices, and refining their contracts to prepare for this new reality, while trade associations across both the creative and tech industries are heavily lobbying to advocate for changes in the law.
Synthetic Voices and Image Rights: A New Battleground
The ability of AI to generate synthetic voices that mimic real artists adds a new layer of legal complexity. In the music industry, an artist’s voice is more than a sound; it’s a unique, monetizable asset and synthetic voices potentially blur the boundaries of intellectual property rights. They raise questions about ownership, identity and originality that our current legal structures aren’t fully equipped to answer.
This issue is particularly significant in the UK, where image rights protections are less robust, leaving artists vulnerable to unauthorized imitation of their voices by AI system users. Further complicating this dynamic is the challenge faced by digital service providers inundated with fake uploads daily.
This concern mirrors the issues being raised in the recent actors’ strike in the United States, where actors are objecting over studios seeking to exploit their image and likeness for future projects without proper compensation. Given the pace and cost-efficiency of music production compared to TV and film, the issue is even more pronounced in music. The swift creation and upload of new tracks using synthesized artist voices can overwhelm artists’ ability to monitor or prevent misuse.
Digital service providers are already trying to find a technological solution, with Deezer announcing that it has developed AI-powered technology to identify deepfake vocals on uploaded tracks and let labels and artists decide what to do with them. Could this be the start of a Content ID-style approach where artists and labels can decide whether to track, monetize or block such content? This is complicated by a couple of factors: (1) it may become more difficult to detect AI-generated works, particularly on UGC platforms; and (2) whether an AI-generated work or use of deep-fake vocals infringes third party rights needs to be considered on a case-by-case basis and may vary by territory, meaning a blanket approach to the treatment of AI content will not work and platforms will need to take care not to over-police.
AI in Music: Anticipating the Future
Though the future is still uncertain, it’s clear that AI has significant potential to reshape the music landscape. The industry needs to carefully navigate the balance of opportunities and risks. Priorities must include respecting artists’ rights, developing a clear legal framework, and fostering creativity. With the right balance, the interplay of AI and music can create a new paradigm in technological and creative innovation.
Just as the advent of digital recording and online streaming revolutionized the industry in the past, AI could be the next big disruptor, dramatically changing the way we produce, distribute, and consume music. As we move forward, it’s crucial that we embrace these changes while ensuring a fair and equitable environment for all stakeholders in the music industry.
Music Business Worldwide