[STUDENT ESSAY] How Platform Regulation and the Misinterpretation of Law Reduce Creativity Online

By Laure Niclot

The Digital, Governance and Sovereignty Chair will now publish, on a regular basis, the finest essays and papers written by Sciences Po students in the course of their studies.

This blogpost features the essay written by a second-year student at Sciences Po (Reims campus), as part of the course taught by Rachel Griffin and entitled ‘The Law & Politics of Social Media’.

How Platform Regulation and the Misinterpretation of Law Reduce Creativity Online

Social media are often said to have revolutionised our lives. In the realm of creation, they have enabled thousands of content creators to share their work with a potentially wide audience, without going through the obstacles of finding a publisher or producer or having to conform to certain standards imposed by the industry. In this new world of social media, there is no longer a monopoly of a few companies deciding who gets to have their work shown to the world: everyone is supposedly free to do so.

 The question we may ask today is about the relationship between social media and creative workers. The latter will be understood as all the people who engage in creative activities of any sort and publish their content online. Such a category covers a very wide range of creative content, from original music and writing to remixes, but also vlogs and lifestyle content. If for some creators, sharing their work online is only a way to extend their audiences or be part of a creative community, for some the content they create would have no meaning were it not published on the internet. Some even make most of their revenue based on their online activities.

Based on these considerations, it is relevant to ask ourselves whether or not social media actually benefits those creative workers. I will make the case that platforms’ content moderation and curation mechanisms present several obstacles to their activities, but also that policies and legal confusion brings about chilling effects that largely undermine both the quality and the quantity of work created and displayed online.

Content Moderation & Curation as a Challenge to Creative Workers’ activities

Bias and discrimination in enforcement of moderation policies

Though social media platforms are often regarded as temples of freedom and free speech, it is not always the case. Indeed, to ensure compliance with the law and avoid any liability but also to make sure that users have an enjoyable experience, platforms enforce moderation policies on the content being posted. Though the point is not to highlight the complexities of such procedures, it is relevant to underline that they have real consequences on creative workers.

AlgorithmWatch has indeed demonstrated that an algorithm used by Google to rate text by potential toxicity has a strong bias against several protected categories. For instance, the use of the expression “as a black man” gives out a 46.2% toxicity score when the score is only 5.3% for the expression “as a French man”. This is explained by the fact that such algorithms are trained on a given set of data that led it to believe that “black man” or “homosexual” are insults because they are very often used as such. The limits of algorithmic content moderation are thus clear: without the ability to understand the context of the comments, it becomes rapidly discriminatory and impedes certain categories of people from having a voice on the platforms. 

A paper by Gretchen Faust illustrates very vividly the consequences it has on creative workers. She pays attention to female artists using Instagram to publish content (mainly photographs) who had to face the deletion of their posts. One striking example is Rupi Kaur, who published a series of photographs relating to menstruation where she represented herself (fully clothed) with a red dot to illustrate her period. The picture was deleted again and again by the platform, without any explanations. What makes this case interesting, as highlighted by Faust, is that Instagram policy mentions that it removes “violent, nude, partially nude, discriminatory, unlawful, infringing, hateful, pornographic or sexually suggestive” content, and that nothing here would automatically qualify Rupi Kaur’s work as removable – and yet it was, proving that the enforcement of moderation policies also entails a patriarchal – or at least discriminatory – dimension.

On a wider scale, the paper also highlights the fact that some platforms’ rules are actually far more conservative and restrictive than the jurisdictions in which they apply. This is for example the case with Instagram, which deletes pictures including female breasts (and not male ones), whereas going around town topless is lawful in many jurisdictions for both men and women (as exemplified in New York by the Free the Nipple movement, where an activist strolled around the streets to denounce this inconsistency).

Platforms’ content moderation policies and enforcement methods are thus sometimes real obstacles to the total liberty of content creators.

Curation policies as an illustration of the differentiated interests of content creators and platforms

Curation methods are the main reason we use social media: they rank and prioritise content so that we do not end up faced with the sea of content posted every day. They are also a way for platforms to ensure that certain content gets the most attention, which makes their business model based on ad revenue feasible. This paragraph’s main focus is on the monetisation system of YouTube (through which content creators get paid when an ad is placed before their videos) as an example of curation policies (pushing forward certain types of content so they get more views but also become more attractive for advertisers).

The entire ad system is fairly complex and opaque, as is the platform’s curation algorithm of the platform. But certain studies, made by experts or content creators, have shown what seems to be a correlation between certain topics and demonetization of the content. One study focused particularly on queer content – almost all vocabulary linked to the queer community seemed to trigger demonetisation of videos. This truly matters, as even if queer terms do not lead to deletion of the content (as was the case for the content examined in the previous section), their visibility and economic profitability are greatly reduced. This really shows that platforms have other motives in shaping their policies than allowing for complete liberty of creation: they also need to put forward content that will be deemed uncontroversial enough so that brands will be willing to be associated with it by putting an ad before it.

            Other examples of the limits curation imposes on creative workers are to be found in the workers who cannot sustain their livelihoods because their videos often get demonetised and have to put in place other systems of remuneration, for example through Tipee, a platform that allows creators to put in place systems of monthly donations in exchange for previews of their videos or other forms of rewards. Two French YouTubers illustrate this point well: the owners of the channels Les Revues du Monde and C’est Une Autre Histoire, both run by female historians who often address female history (with videos on the history of abortion or of periods) being pushed to resort to donation system to sustain their business models because their content often gets demonetised due to its “controversial nature”.

Curation methods, by deciding whose and what type of content can get both attention and ad revenue, are thus symptomatic of a system where creative workers are vulnerable to the economic interests of platforms.

How legal and algorithmic confusion create obstacles to creativity

Hyper-awareness of content creators of the variations of the algorithm

Beyond the fact that certain content is harder to live from or to receive attention on because of the curation policies, it must also be highlightedthat if certain creative workers carry on with this type of “controversial” content, it is certainly a huge disincentive for most. Influencers are known to try to figure out how the algorithm works, which type of content and format gets the most attention or the better recommendations on the platform, and to shape what they create according to those norms rather than what they would have created otherwise. As a consequence, we can say that when creative workers depend on the platforms to obtain their main source of income, it creates real restraints on their creativity as they have to adapt to the constraints of an unknown algorithm.

This creates tendencies to “gam[e] the system”, meaning having strategies adapted to each platform to make sure the content posted gets attention. Those practices are also deeply linked to the confusion that surrounds algorithm and curation policies: the lack of clear guidelines of what is preferred by the platforms pushes creative workers to spend countless energy on guessing what will work rather than actually creating. This hyper-awareness of the constraints linked to the promotion of their work is deeply linked to the confusion that fuels those unsaid rules.

This means that rather than producing the total liberation of creativity in an online environment free of typical reputational or economic constraints, social media have enhanced a model where a main priority for creative workers is the thought of fitting the current expectations of the platforms. Which, if compared with the initial state of the creative industry before the appearance of social media, does not seem to be much different.

Copyright issues: how confusion in legal understanding and frameworks leads to irrelevant constraints on creation

Another relevant case where we see how the confusion surrounding the enforced rules prevent creative workers from being fully free is the one of remix practices: when creators take an already existing piece of work and “remix” it to produce something partially original. Examples of such practices include fanfiction, music remixes, drawings based off existing characters or figures etc. In this context, the issue of copyright is more important than that of curation policies, since there is little financial pursuit in such practices. But the same mechanisms of opaqueness apply.

 Within such communities, copyright is the central issue on both ends: to protect themselves from infringing the law based on the work they produce but also to protect such works from being taken by other creators. A study by Fiesler et al. produced a qualitative and quantitative analysis of comments on such forums that revealed an overall fear of law infringement, whether they be recognised to have stolen from others or that others have plagiarised them. Some creators might thus decide to leave the platforms they deem uncertain (because they overblock for instance) due to this confusion over what can and cannot be done.

 Because of this legal gap, either felt or real, social norms also arise, that might lead to stricter rules on copyrights. For instance, the paper mentions the policy implemented by the fanfiction platform “Twisting the Hellmouth” on the case of creators wishing to do a fanfiction based on another fanfiction, in which you have to ask permission from the original author (of the fanfiction) to write such a story, knowing that the author could always change their mind later on. This mere codification of a social rule originating from a creative community shows that confusion around a legal scheme leads to stricter rather than more permissive policy frameworks. A similar parallel could be drawn with what we have mentioned on the nudity policies of Instagram – showing once again that social media are far from being havens of free creation.

Overall, this lack of information leads to chilling effects, meaning that out of fear of legal consequences (in the case of copyrights) or reputation spikes, creative workers refrain from creating pieces they would have otherwise created.

To conclude, this blog post has suggested that the moderation and curation policies of social media platforms are not tools of freedom for creative workers, but rather that they created points of restraint for their contents. We have also underlined that these obstacles are also partly due to a less-than-clear understanding of both the legal system surrounding creative practices and the policies of the platforms that lead content creators to adopt behaviours they would not have otherwise adopted, such as adapting their creations to the perceived expectations of algorithms.

It seems to be the case that social media, though they have indeed benefited a part of the creative communities by allowing influxes of shared content to grow exponentially, have not been able to provide the most adapted place for creative workers. The economic constraints of the offline world are still very much present in the relationship that unites two diverging stakeholders: social media platforms and creative workers.

---

Laure Niclot is a second-year student on the Reims campus of Sciences Po. Particularly interested in legal matters and politics, she likes to look at how law and policies interact with our behaviour.