Embracing AI while Protecting IP

Share on:

April 29, 2024

As artificial intelligence continues to be intertwined with our daily lives, its impact on music is increasingly prevalent. From initial composition to the final phases of distribution and marketing, the entire music industry value chain is being impacted. AI applications in music range from simple recommendation systems and supportive tools to advanced generative AI models. Among the most prominent use cases are text-to-music generators like Suno, automated mixing and mastering solutions such as Roex, and vocal cloning apps like Kits.

The duality of AI in the music industry

AI in the music industry can be seen as a double-edged sword – an opportunity for increased creativity and efficiency versus a serious threat to IP & copyright protection. While AI holds promise, the majority believe the risks outweigh its opportunities. The inevitability of AI’s future necessitates an urgent industry commitment to establishing mechanisms that guarantee responsible and ethical utilisation of this technology in order for the opportunities to truly shine. 

AI music challenges

At the heart of the issue is whether and how generative AI is permitted to use copyrighted content as input for AI model training. Unfortunately, copyright law was not designed for AI, so no clear answer to this question currently exists. Consequently, the industry currently faces four main challenges: preventing copyright infringement and unethical use of IP, identifying AI-generated music, and establishing mechanisms for fair compensation.

  • There are no mechanisms to protect IP & prevent copyright infringement.
    AI models analyse large amounts of musical data from every corner of the internet. They then discover patterns, learn from this data, and generate new sounds. AI models can thus be trained on an artist’s IP without their consent or authorisation. Since copyright law was not defined with AI in mind, uncertainty exists if IP is even protected from AI in a legal sense.
  • There are no mechanisms for fair compensation.
    IP rightsholders seldom receive compensation when AI-generated music is derived from their IP. Following the “Heart on My Sleeve” case, many have taken legal action regarding copyright infringement and lack of compensation for rightsholders. Comparable lawsuits involving the legal implications of AI are currently making their way through courtrooms, and this issue is expected to escalate as the utilisation of AI increases.
  • There are no mechanisms to identify AI-generated music.
    Whenever an AI model generates an output, no labelling mechanism identifies it as AI-generated content. This means there currently exists no way to distinguish AI-generated music from human-made music.
  • There are no mechanisms to prevent unethical use.
    As AI-generated music or voice models reach a stage where they are almost indistinguishable from their human counterparts, more ethical considerations emerge regarding unethical use cases – such as spreading hate speech using someone’s voice clone. No mechanisms exist to penalise those who spread unauthorised content using an artist’s voice clone.

How does blockchain empower AI while protecting IP?

To ensure fair and ethical use of AI music, it is crucial to establish a system that provides efficient and transparent mechanisms for IP ownership, licensing, distribution, and provenance. Blockchain technology can reform how rights are managed, licensed, and compensated – bringing efficiency and clarity to the sector.

The transparent, decentralised, and immutable nature of blockchain, combined with tokenisation – a process of converting the rights of an asset into a digital token on a blockchain – enables musicians and rights holders to establish undisputed digital ownership of their IP. While much of blockchain’s interest revolves around finance, its fundamental purpose is verifying authenticity. This tokenisation process, usually in the form of a non-fungible token (NFT), means that all information surrounding the IP is digitally preserved about who created what, who owns what, and where it is used. 

Rightsholders can then programme authorisation and usage rights into their tokenised IP. This allows smart contracts to perform certain automated processes, such as transferring ownership, managing royalty distribution, or licensing copyrights for AI model training. 

When a rightsholder chooses to license their tokenised IP for AI model training, blockchain can ensure complete provenance of the process. Since the process is traceable from the moment of IP registration, through AI model training, and ultimately to consumption – rightsholders could automatically earn royalties whenever an AI model trains or produces content that includes their IP.

A transparent tokenised process ultimately empowers both sides of the coin. Rightsholders gain the ability to control and monetise their IP, while companies developing AI models avoid legal disputes. In this way, blockchain technology and tokenisation provide a transparent foundation for establishing fair and ethical ownership, licensing, distribution, and monetisation mechanisms for AI-generated music.


Brought to you by the Web3 Music Association (“W3M”) – a non-profit entity with the goal of orchestrating innovation in the music industry. Its mission is to educate music industry professionals, support their digital transformation, and bring them together to collaboratively develop innovative use cases. Created from an extensive three-year collaboration, the association is a lead contributor to the Music Protocol – a dedicated blockchain for intellectual property registration, management and monetisation.

Contributor

Sergio Mottola – Web3 Music Association President and CEO

Follow