![]() |
| Photo by New York Said on Unsplash |
Copyright infringement claims focused on output, not training AI models
While the complaint explains that Suno’s training dataset involves all ‘music files of reasonable quality that are accessible on the Internet’ and user uploads of music files – and thus – concluding that its model has memorized many of these works; its infringement argument is limited to output. In perhaps a global first, SOCAN argues that Suno have infringed and authorised the infringement of SOCAN’s performing rights pursuant to sections 3(1)(f), 2.4 (1.1.) and 27(1) of the Canadian Copyright Act 1985.
While the complaint explains that Suno’s training dataset involves all ‘music files of reasonable quality that are accessible on the Internet’ and user uploads of music files – and thus – concluding that its model has memorized many of these works; its infringement argument is limited to output. In perhaps a global first, SOCAN argues that Suno have infringed and authorised the infringement of SOCAN’s performing rights pursuant to sections 3(1)(f), 2.4 (1.1.) and 27(1) of the Canadian Copyright Act 1985.
The argument is two-fold: (1) Suno encourages users to prompt the generation of output that is identical or substantially similar to the listed works; and (2) Suno allows these outputs to be streamed on demand by both the user and potentially other members of the public.
SOCAN’s complaint identifies similarities in lyrics, melody, and harmony between the listed works and Suno outputs. Some involve generating lyrics in another language (e.g. Korean version of “Sk8er Boi”), while others prompt a different musical style (e.g. 1990s grunge rock version of “Big Yellow Taxi”). SOCAN maintains that this output is substantially similar to the listed works including those titled ‘remixes’. SOCAN cites user comments that remark on the similarity or even praise the user for near-identical similarity with the original work as evidence that these similarities are immediately recognizable to a layperson of the intended audience for the works (perhaps reminding us of Pelham II pastiche criterion… see IPKat here). It also alleges that that some outputs are posted on third-party streaming services (e.g. YouTube) for commercial gain.
SOCAN alleges that Suno makes the listed works available on demand by offering to generate substantially similar outputs. Once generated, they are available for on-demand streaming through embedded links and communicated to the public through Suno’s platform. SOCAN further argue that Suno authorises users to make the listed works available for on-demand streaming on third-party streaming services or websites, and to use the ‘remix’ or ‘remix/edit’ function to existing outputs.
Lastly, SOCAN describes Suno as encouraging, or at least being indifferent to, the infringing behaviour or its users. Instead of taking proactive steps to operate in a lawful manner SOCAN alleges that it hides behind its terms of use. Alongside the recent introduction of user guardrails, SOCAN maintains that infringement is a foreseeable and intended consequence of Suno’s core business model that entitles it to punitive and exemplary damages.
Comment
If this case goes ahead, it will undoubtedly be one for copyright aficionados worldwide as it has a little bit of everything – indeed it would make a great essay question for promising copyright students (Liverpool Law School students, you’ve been warned!) Geeky copyright antics aside, the complaint engages recent challenging policy discussion concerning copyright platform liability and responsibility, fundamental rights and, likely once Suno responds, the weighing of interests, specifically user and public interests through exceptions and limitations.
When describing the nature of the allegedly infringing activities, SOCAN references the Canadian originality standard of skill and judgement (CCH Canadian Ltd. v. Law Society of Upper Canada). It explains that Suno is ‘shockingly efficient at imitating human-created music’ and ‘does so at blazing speed on a massive scale’. This technical ability is ‘central to Suno’s commercial offering’ as it ‘commercially exploits musical outputs that replicate the original expression embodied in existing music’. SOCAN states that Suno ‘has over 2 million paid users, collects over $300 million USD in annual recurring revenue, and generates an estimated 7 million musical outputs per day'.
Although SOCAN juxtaposes the creative labour of authors against Suno’s copying at scale, for this Kat, questions arise over user remixes and copyright exceptions (e.g. Canada’s non-commercial user-generated content exception pursuant to s29.21 of the Copyright Act). There is certainly an argument that Suno’s platform ‘diverts attention, listening time, and revenue away from music creators, music publishers and other rightsholders while it profits from the infringement of their rights’. Indeed the complaint refers to market dilution theory (discussed in Kadrey v. Meta Platforms – IPKat here and here), where human-created works compete with AI-generated outputs, to emphasise the devastating consequences on human creativity.
Less clear, however, is the position of user creativity, particularly when it occurs through commercialised digital environments. Similar arguments, of course, were made regarding Article 17 of the Digital Single Market Directive. In a similar vein, SOCAN explains that it is impossible to identify the full scope of infringing activity given the exponential amount of outputs made available to the public, and those made to individual users. Suno, they allege, knows the full extent and continues, brazenly.
In this Kat’s view, this creates a perfect storm where legislative intervention is challenging, but necessary. How can the contours of the copyright system be respected in a manner that supports all forms of creativity without destroying its very foundations? It also leaves a wider question over the role of copyright, culture and technology as the complaint makes clear that:
"the authors of the Listed Works represent a wide variety of cultures and communities in Canada and range from up-and-coming talents to some of the country’s most internationally renowned songwriters and composers."There is something inspiring in SOCAN submitting a list of works comprising Canadian musical heritage, both young and old. It mirrors Koda’s pending litigation against Suno in which Koda’s legal and policy director, Nicky Trebbien has stated that:
"It is completely unacceptable that it takes a lawsuit to make Suno and similar AI services pay for the music they use. If we want to continue having talented artists creating new Danish music in the future, we must protect our music industry and not leave it to cynical tech companies and their algorithms to shape and tell our story."For this Kat, it affirms her view that the socio-cultural value of AI model training sources and subsequently generated content far exceeds the borders of the copyright system. It impacts norms related to humanity’s shared cultural heritage that she terms ‘artistic data’. Such an approach requires viewing creativity as an expanding set of ongoing creative communication between authors, users and the public. So when this communication provides a foundation for platforms like Suno to exist and significantly, profit, artistic data sovereignty allows a collective redress mechanism without instilling more distrust of the copyright system or impeding future user creativity (see further‘Invisible ink: Artistic data and the erosion of creative cultural agency’ and the forthcoming book Creative Reuse).
The SOCAN v Suno complaint is a strong first step to begin to account for this value which humanity, more than ever, depends upon.
Do you want to reuse the IPKat content? Please refer to our
Policies section.
If you have any queries or requests for permission, please get in touch with the IPKat team.

Comments
Post a Comment
All comments must be moderated by a member of the IPKat team before they appear on the blog. Comments will not be allowed if the contravene the IPKat policy that readers' comments should not be obscene or defamatory; they should not consist of ad hominem attacks on members of the blog team or other comment-posters and they should make a constructive contribution to the discussion of the post on which they purport to comment.
It is also the IPKat policy that comments should not be made completely anonymously, and users should use a consistent name or pseudonym (which should not itself be defamatory or obscene, or that of another real person), either in the "identity" field, or at the beginning of the comment. Current practice is to, however, allow a limited number of comments that contravene this policy, provided that the comment has a high degree of relevance and the comment chain does not become too difficult to follow.
Learn more here: http://ipkitten.blogspot.com/p/want-to-complain.html