Post by Aline Iramina, PhD student in Law at University of Glasgow
On Sunday, 28 August 2022, Anitta made history as the first Brazilian winner and performer in a MTV Video Music Awards (VMAs). However, before that, she was already breaking other records. For those who do not know, on 24 March 2022, Anitta was the first Brazilian and Latin American solo artist to achieve the top of Spotify’s global 50 chart with her single “Envolver”, which received more than 6.4 million streams in one day, with 4.1 million streams coming from Brazil. At the time, this achievement of a Brazilian artist made headlines across the country. Besides her music choreography trending on TikTok, one of the reasons behind this success was a huge mobilisation of her fans and Brazilian artists and influencers, especially on social media, encouraging people to listen to her music. However, this was not the only reason for her success, and this is why Anitta’s Spotify record is a good case to introduce the concept of “Copyright Governance by and of Algorithms” in this blog post.
In an article on Rest of World, Anitta’s superfans were accused of gaming Spotify’s algorithms. In this article, they accused her fandom (“Anitters”) of potentially “violating” Spotify terms and conditions, when they used Twitter to encourage other fans to boost “Envolver” numbers, creating playlists and explaining exactly how they should listen to it. For example, on Twitter, in a post of one of Anitta’s fan accounts, they explained how people should use different Spotify’s accounts and remember to change their accounts after twenty streams so they could be counted separately. This is because the platform does not count past twenty streams when the track is played on repeat from a single account, since they consider this to be bots used to inflate a track’s streams. This is not allowed according to Spotify and other music streaming platforms’ rules.
However, Anitta’s fans were definitely not the first to use algorithms in their artist’s favour. This actually happens quite frequently, especially for artists with huge fanbases. We can easily find similar practices in other fandoms, such as those of BTS (“Army”) and Justin Bieber (“Beliebers”) on Twitter. As a Brazilian journalist said, if Anitta, her team and fans achieved the top spot, it is because they were able to use the platforms’ systems and business model better than other big global artists. So, as mentioned in the Musically’s article about this case, what is crucial here is to understand the importance and role of fans to the global music streaming economy. More importantly, it is necessary to analyse whether new platforms and chart bodies’ rules will be necessary to govern what is considered acceptable, especially when record companies and artists can engage with fans and provide them with incentives to push their music up in the charts. However, the debate is not only about how platforms, labels, and chart bodies should deal with cases like Anitta’s. It is how all the main actors deal with this, including governments, within a copyright governance system.
When we talk about governance, there are many aspects that should be discussed. As argued by Lawrence Lessig and his “Code is Law” theory, there are four main governance mechanisms that influence individuals’ behaviour: the law, social norms, the market, and architecture (the code on the internet). Copyright governance involves all these aspects. If boosting streams by gaming recommendation algorithms and creating innumerous playlists was not socially acceptable, artists and fans like Anitta and “Annitters” would feel, at least, less comfortable about adopting this kind of practice. However, this is not the case since these practices are very common between artists’ fanbases worldwide. Platforms, labels, and chart bodies’ decisions are certainly important to determine how the music industry works. Spotify’s user guidelines stipulate that those who use bots or any other automated systems to artificially increase their number of streams should have the access to their accounts suspended. Labels negotiate with platforms to boost their artists’ streams by getting their music promoted on the digital service, using their marketing tools, such as Spotify’s Marquee and Discovery Mode. We should remember that, since it was launched in 2020, the pay to play Discovery Mode has been questioned especially by independent artists and labels. Spotify argues that this tool allows musicians to be discovered, by adding a signal to the algorithms that determine personalised recommendations. In this case, in exchange for accepting a lower royalty rate, artists have their music suggested by Spotify’s recommender algorithms to listeners that otherwise would probably not discover them. Musicians have questioned whether this would not be a form of payola, which was considered illegal within traditional media, benefiting big labels and artists who can afford to pay for this tool. Not only that, as mentioned by Artists Rights Alliance in an article they wrote for the Rolling Stone, one of their concerns is that they might end up paying for this new tool and gaining no extra exposure for it, since “if everyone is boosted, nobody is”. They demand more transparency about Spotify’s algorithms and data processing, so they can make informed decisions and pursue their commercial interests when using the platform to distribute their music. To better understand how algorithms work is a real concern within the music industry. Algorithms are used as a tool of governance influencing and shaping users’ listening habits and, consequently, artists’ success in achieving a larger audience and a fairer remuneration. Hence the requests in different parts of the world to governments intervene to regulate algorithms’ use by digital service providers.
Therefore, the discussion should not only be about a “copyright governance by algorithms”, but also about a “copyright governance of algorithms”. As explained by Latzer and Just (2020), while the “governance by algorithms” paradigm focuses on the algorithmic systems and their economic and social impact on individuals and society, the “governance of algorithms” considers the different forms to “control, shape, and regulate algorithms and the effects that results from their processing, learning/adaptation, and/or decision-making powers”. When we discuss copyright governance of algorithms, the idea is to analyse the new rules adopted by different copyright actors, in particular the state, to govern the use of algorithms and their impact on creators, copyright users and the creative industries. In the music industry, besides the state, we saw, while discussing Anitta’s case, that online platforms, labels, chart bodies, even artists and users have the power to control or shape algorithms and their effects. In my PhD thesis research, I try to focus on all these aspects, but with a greater focus on states’ role in this governance system and the new algorithmic transparency rules they have adopted.
Within copyright law and creative industries, the focus for many years has been on content moderation systems, as the debates surrounding the approval of article 17 Copyright Digital Single Market Directive (CDSMD) in the EU have shown. However, in recent years, recommendation systems have gained more attention from copyright stakeholders. In the UK, the “Economics of Music Streaming” inquiry conducted by the UK Parliament includes many concerns presented by stakeholders about the use of algorithms by music streaming services, especially in relation to the lack of transparency and potential bias in the employment of recommendation algorithms. Many artists requested the “oversight of platforms so algorithms are not biased”. In their contributions, we can see the difficulties faced by new and independent artists to make Spotify’s recommendations systems work in their favour. If they do not have a significant fanbase like Anitta and other popular artists, or a record label to support them, they struggle to have their music recommended on these platforms. Moreover, they argue they do not have enough information about how these algorithms work in practice and whether they favour some artists to the detriment of others. The DCMS Select Committee’s final report mentions how important algorithms are to streaming services and the necessity for more clarity about their influence on music consumption and the level of oversight that already exists. In Europe, the adoption of a new platform regulation, the Digital Services Act, will introduce new transparency rules for the use of algorithms that will have a great impact on online platforms, such as YouTube and Spotify. In the case of Spotify, as a hybrid platform, this will depend on how the DSA rules will be applied, since only the Spotify podcast service would fit in the category of an” intermediary service”. For now, at least in the EU, artists will probably rely on article 19 CDSMD to request more information from music streaming services about the influence of recommendations algorithms on their music streams and their remuneration. If they will get the information they want, we must wait to see.
When analysing the new laws and regulation proposals, we see that most states have focussed on transparency and accountability rules to regulate the use of algorithms. Although most of these rules were not conceived initially to address creators’ concerns on platforms’ algorithms, they also apply to them. The idea behind these new transparency and accountability rules is that first we should understand how algorithms work and their real impact on individuals and the society, so we can properly regulate them. Despite the criticisms to this approach and the limitations of transparency, it is an important first step to regulate algorithms. To conclude, it is interesting to mention that in a recent talk on Codex Future Law 2022, referring to his “Code is Law” theory, Lessig observed that now ‘business models eat law’, when asked about Facebook´s metaverse and platforms’ power. For him, the discussion is not more about how technology (the code) governs individuals and society, but how big tech companies’ business models override the law of states; how their business models focus on digital engagement despite their effects on democracy and individuals’ rights.
It will be interesting to observe how states plan to overcome this situation and whether they have the power to regulate these big tech companies in the following years, but this is another story. For now, in the EU, the new platform regulation, in particular the DSA and the DMA, signal the direction they want to take. This includes new transparency rules to regulate the use of both content moderation and recommender algorithms by online platforms. Regarding Anitta and Spotify, cases like this will probably continue to be governed by private actors and their own rules, as it supposed to be. This does not mean that the new regulation will not impact music artists and creators that rely on online platforms to promote and distribute their works.