Who Owns the Sound? AI in Music and the Legal Landscape

AI-generated music is challenging traditional copyright frameworks, raising questions about ownership, legal compliance, and the balance between AI innovation and protecting artists’ creative rights.

SwissCognitive Guest Blogger: Shivi Gupta – “Who Owns the Sound? AI in Music and the Legal Landscape”

Attending a live gig, enjoying the music from your favourite artist or band? What if it can come to your couch, with the feeling that they are performing right there in front of you? But hey, who is producing the music? Is it AL or Al

The creation is revered, but more than the creation, the creators are worshipped. Recently, Sony, Universal, and Warner have sued Suno and Udio (GenAI music startups), claiming copyright infringement in training models, to protect the artists affiliated with these giants.

Major record labels are protecting their clients, the artists, the great ones who produce music that can rarely be replicated. But in the day and age of generative AI or (GPT Generative Pretrained transformer), music is also replicated by machine learning algorithms to make songs sound like the original creators.

As one of the popular web3 music websites Unchainedmusic.io wrote in their article “Deepfake vocal synthesizers, an innovation in AI technology, can make a singer’s voice sound like a famous artist. Under English and EU law, it is unlikely that a style of singing, whether generated through deep learning, AI or vocal imitation, is protectable by copyright. However, other forms of intellectual property, such as passing off, may be relevant in some jurisdictions.”

There is no common universal law against intellectual property, and most countries have their own rules, copyrights, patents. Any commercial use begets a request or a permission from the creator who owns the intellectual property of their voice.

Problem:

All music can be created eclectically with different styles, lyrics and genres. GenAI music might saturate the market with more and more music generated by machine learning algorithms.

Possibility:

Music lovers will rely on humans creating songs as it has the emotional factor, the timber, tone, pitch, stretch, diction, accent are some of the unique human characteristics which helps us being empathetic and understanding of the singer’s mindset.

Probability:

These AI created songs will be used by ad companies and video editors to feature a product or sell a service with an attractive UX.

Musicians will continue creating great records and go on tours, and fill stadiums.

Editors, marketers, sales representatives will use GenAI music in elevators, advertisements, branding, showcase of their products and services. The GenAI music will complement the product *-as-a-service.

Proposed Solution:

Follow rules created by the countries in which these AI tools are used. For classical music the law states that as mentioned by edwardslaw.ca “it covers original literary, dramatic, musical, and artistic works of authorship. This is during the lifetime of the author, the remainder of the calendar year in which the author dies, plus 70 additional years (the Canadian copyright lifespan recently increased from 50 to 70 years in June of 2022). Once this term expires, the work becomes public domain. “ So works from Beethoven, Mozart et al. can be performed in public without permission or paying a fee – royalty free. So any music which has been recorded prior to 1974 can be used since it has entered the public domain, but if you have the London symphony orchestra uploaded their recording of Beethoven’s symphony number 5, one can’t use that without the permission from the orchestra.

For example this particular youtube video can’t be reused without BBC’s permission:

More on copyright of voices: “According to Herndon, much of vocal mimicry comes down to personality rights. “You cannot copyright a voice, but an artist retains exclusive commercial rights to their name and you cannot pass off a song as coming from them without their consent,” she wrote in a recent Twitter thread, citing previous legal cases related to vocal impersonation.”

More about ethics in AI.

About the Author:

Shivi Gupta is a passionate data scientist and full-stack developer, working in the industry for over a decade. An AI expert navigating through the world of real vs generated content. With a focus on ethics , he creates websites, mobile applications, chatbots all powered by AI.