Uploaders' perceptions of the German implementation of the EU copyright reform and their preferences for copyright regulation

Introduction

Video platforms have fundamentally changed the traditional division of roles into that of producer and user. Anyone can now easily switch between these roles (Cha, 2014; Ding et al., 2011), and media content producers further lose control of how their content is used, leading to copyright conflicts. Those conflicts occur in “a triangular relationship” (Shepard, 2014, p. 731). Uploaders want to use as much content as possible, platforms favour maintaining a “neutral” position and rights holders (e.g., musicians or media companies) prefer to engage platforms as debtors for liabilities (Holznagel, 2020). Under the previous legal situation in the European Union, the standard case was that the platforms were excluded from liability for their users’ copyright violations, with exceptions set by the case law of the European Court of Justice (ECJ) (Frey, 2022, p. 97). However, the European Union’s new Directive on Copyright in the Digital Single Market (DSM Directive) incorporates a direct liability of platforms for giving access to copyright infringements constituted by their users as the standard case (Gielen & Tissen, 2019). The platforms can be exculpated from this liability by concluding license agreements, or by making sure that questionable content is inaccessible (Art. 17 para. 4, DSM Directive), which means the de facto use of upload filters (Spindler, 2019, p. 286).

The DSM Directive has caused protests across Europe. Content producers were convinced that it would mark “the end of the Internet”, endanger freedom of expression and lead to overblocking (Brost et al., 2019). Moreover, critics asserted that the DSM Directive – despite protests and controversial public discussion – was voted on “behind closed doors” (Sagatz, 2019), and that the demands of the (often weakly organised) content creators were not considered in the legislative process. Our study takes up this point and asks how content creators on YouTube perceive the current copyright regulations, the implementation of the reform in Germany and what further needs for regulation they see.

That said, the contributions of our study are threefold. First, previous literature focuses on different forms of content creation activities (further called usage practises), their popularity, and their acceptance in the community (Ding et al., 2011; Erickson et al., 2013; Hui, 2021; Werner, 2012; Xie et al., 2011; Xu et al., 2016). Second, previous literature has discussed uploaders’ evaluations of copyright regulatory frameworks; however, considering the low impact of state regulation, studies have focused on private copyright-related content moderation systems, such as YouTube’s Content ID (Gray & Suzor, 2020; Hui, 2021; Kaye & Gray, 2021). Third, there is a lack of research dealing with current changes in the copyright context that takes into consideration the (missing) perspective of active content creators. Linked to that, our study fills a gap in public discussions of the Act on the Copyright Liability of Online Content Sharing Service Providers (Urheberrechts-Diensteanbieter-Gesetz, UrhDaG), which transposes the DSM Directive into national law1 by giving a voice to users who actively engage in the participatory culture on video platforms, but did not take part in the process of submitting public comments during the consultation procedure on the draft of said law. Finally, despite the focus on Germany, our study contributes to the global copyright discourse by identifying potential challenges in copyright regulation on video platforms, providing possible solutions and insights for future research.

After reviewing previous literature, in the context of the mentioned contributions, we conducted a qualitative content analysis of comments on the draft of the UrhDaG (n = 17) and additional qualitative interviews (n = 19) with active uploaders, aiming to fill the gap in the public discourse. Based on the results of our analysis, we discuss the perceptions and demands of uploaders on copyright regulation with regard to the implications for further copyright regulations.

Determinants of the perception and evaluation of copyright law and its future requirements

Different types of content uploaders, their usage practises and the relevance of these practises for potential copyright infringements

The majority of users on video platforms only use them passively to watch content. Only 1 percent of YouTube users upload videos and can thus be described as “active uploaders” (Ding et al., 2011, p. 361). As the aim of our study is to obtain a deeper knowledge of the perceptions and needs of these active users regarding copyright regulations, we concentrate on this group of uploaders.

The uploaders on video platforms vary regarding their degree of professionalisation (e.g., individuals vs. media companies). In this vein, the degree of professionalisation of uploaders can be conceptualised as a combination of their degree of organisation and the outreach they gain with their channels, that is, the number of subscribers they have (Marek, 2013, p. 48; Zabel et al., 2017, pp. 132–133). Both aspects have a direct influence on how uploaders deal with copyright regulations, as they increase the available knowledge regarding copyright regulations, the resources for content production and the available legal advice. The payments YouTube grants to its uploaders are generated by displaying advertisements before, during or after the videos. However, only those uploaders who are part of YouTube’s partner program receive a share of the advertising revenue. The requirement for membership is a minimum of 1,000 subscribers (Google, 2021b).

In addition, there are also differences in the aspect of various usage practises regarding their relevance for potential copyright infringements. Scholars classify the content on YouTube in three main categories: “copied”, “appropriative” and “original” (Hilderbrand, 2007, p. 56). First, the user-copied content includes videos initially produced outside of the platform and uploaded by users without any permission from the rights holders. The second category, appropriative content, includes content that is based on the copyright-protected work of others. In contrast to user-copied content, appropriative video productions provide new but derivative works. For example, the established remix culture on YouTube includes several creative forms of appropriative content, such as political remix videos, parodies, memes and mash-ups (Conti, 2013; Henriksen & Hoelting, 2016; Xu et al., 2016). Lastly, the “original” user-generated content category includes video clips that do not contain any copyright-protected work and are often uploaded by amateur users (Hilderbrand, 2007). Based on the reviewed literature, we consider uploaders’ activity on YouTube (i.e., their usage practises), the type of content they upload and whether their livelihood depends on their YouTube activity (i.e., the financing aspect) as the most crucial aspects for shaping their perceptions and demands of copyright regulations. How the uploaders deal with copyright enforcement, how they evaluate the reform proposal, as well as which further regulatory needs they have depend in turn on their actual knowledge of, and experiences with, the existing regulations.

Uploaders’ knowledge, experience and evaluation of copyright regulations

A leading example of automatic copyright detection technology is YouTube’s Content ID system. Its main aim is to scan user uploads for infringements and enforce copyright claims (Boroughf, 2014, p. 2). The system compares uploaded videos against a database with reference files provided by the rights holders (Google, 2021a). When Content ID identifies a copyright match, rights holders can either block the matched video, monetise it or track its views (Google, 2021a; Kaye & Gray, 2021, p. 2). One of the main reasons for tolerating infringements is that the economic interest of copyright holders can benefit from the popularisation of their content (Boroughf, 2014, p. 10). Gray and Suzor (2020) analysed Content ID removals to better understand copyright enforcement on YouTube and found that certain video content on the platform is more likely to be removed than other content (see also Kaye & Gray, 2021). For instance, film piracy, i.e. copies of full movies uploaded on the platform, are more likely to be removed from YouTube than gameplay streams (Gray & Suzor, 2020, p. 6), although all three usage practises can – depending on the circumstances – violate copyright (Wu, 2008, p. 2).

Furthermore, the option to block content that is considered unlawful gives copyright holders power, which can be misused and lead to restrictions on free speech (Urban et al., 2016, p. 116). One of the main challenges to these private copyright regulations on video platforms is the lack of transparency, which is the reason why some uploaders consider themselves disadvantaged by YouTube’s automatic content moderation and perceive it as unfair (Kaye & Gray, 2021). Uploaders point out that due to the sensitivity of the Content ID system, they estimate a high risk of a claim being made, even when their content is not infringing on copyright (Kaye & Gray, 2021). This is especially serious for professional uploaders who earn their living from YouTube videos and are therefore dependent on the platform’s policy. Uploaders further argue that copyright regulations hinder the participatory YouTube culture (Sganga, 2015).

Because uploaders are directly affected by the copyright policy, our first aim is to examine their knowledge, their experiences and their evaluation of current copyright regulations. Additionally, for the evaluation of upcoming copyright law, it is crucial to analyse whether uploaders believe that the current private regulation is fair for all participating parties. Furthermore, in the context of the German implementation of the copyright reform, we want to analyse how different types of uploaders evaluate the law transposing the DSM Directive into German law, and how effective they consider the new government regulations of the UrhDaG in the light of their experiences with private copyright enforcement. The UrhDaG focuses on mitigating the effects of upload filters, thus avoiding overblocking through procedural mechanisms (Sections 9–12, UrhDaG). These procedural measures include the so-called “de minimis limit” (Section 10), which allows the use of up to 15 seconds of video or sound, 160 characters of text or 125 kilobytes of images, as well as the flagging procedure (Section 11). Here uploaders can mark their content as a use authorised by law. The de minimis limit and the flagging procedure are only applicable if less than half of another person’s work is used. When there is either a minor use or the content has been flagged, the content can be publicly reproduced as an instance of presumed authorised use until the conclusion of the complaint procedure that can be initiated by the rights holder (Section 9). Additionally, especially trustworthy uploaders are equipped with a “red button” (Section 14 para. 4), preventing public reproduction, for cases in which the presumably permitted use could lead to considerable economic damage. In the context of the different sides among copyright conflicts, a central question is whether uploaders consider the new regulations to fairly balance competing interests. Based on content producers’ knowledge and evaluations of the UrhDaG, we finally aim to investigate what demands uploaders have of copyright regulation, and to what extent the current copyright regulation meets their needs.

Methods

Sampling

Based on the two aspects of professionalisation – organisational level and reach – we identified nine different types of uploaders with different degrees of professionalisation (Table 1). We distinguish, regarding the degree of organisation, between strongly organised content producers (e.g., companies, organisations and institutions), moderately organised content producers (e.g., a loose coalition of two or more people who run a channel together) and weakly organised content producers (e.g., individuals). Concerning the number of subscribers, our differentiation is in line with Zabel et al. (2017), between channels with high reach (> 5,000 subscribers), medium reach (501 to 5,000 subscribers) and low reach (< 500 subscribers).

Additionally, in this study, uploaders are defined as people who own a YouTube channel and have at least some activity on YouTube, that is, they have uploaded at least one video during the sampling month (February 2021; Zimmermann et al., 2020).

Procedure and Materials

To investigate the different copyright-related usage practises, and to analyse how different types of uploaders evaluated the new regulations, as well as their demands of copyright law, we combined a qualitative content analysis of submitted comments on the draft of the UrhDaG, including qualitative interviews with uploaders.2 The German legislative process provides interested parties with the opportunity to submit comments on draft laws and thus take part in the lawmaking process. These parties submitted 107 comments on the proposal concerning the German implementation of the copyright reform, and these comments were published on the website of the Federal Ministry. The most comprehensive comments came from established and strongly organised players (e.g., Twitter, Facebook and Google) and primarily shed light on legal details rather than on the social impact of the proposal. Even though there were comments by active uploaders in the consultation procedure, they were underrepresented and lacked detailed (expert) knowledge of the reform proposal. Since we were interested in the perspective of active uploaders, we only selected those that met these criteria (n = 17). To these, we added 19 interviews with moderately and weakly organised uploaders to include all types of uploaders (see Table 1). The interviews were conducted online between April and June 2021. In the selection of interview partners, we attempted to achieve a broad variance across the genres.

To structure the interviews, we developed an interview guide following five dimensions derived from previous literature (see above):

• Financing: YouTube activity as a hobby, or a part-time or full-time job, taking part in and evaluating the YouTube partners program.
• Copyright-related usage practises: the genre and type of content YouTubers upload on the video platform. Based on this information, we categorised the uploaded content according to our scale of lawfulness.
• Knowledge, experience and evaluation of previous copyright regulations: general considerations on copyright when producing and uploading content, personal experience with existing copyright regulations (Content ID) and the perceived necessity and fairness of copyright regulations.
• Knowledge and evaluation of the copyright reform regarding the four concrete implementations: introduction of upload filters, the de minimis limit, the flagging procedure and the red button.
• Demands of copyright regulation: needs and requirements of copyright regulation.

After the interviews were conducted, they were transcribed manually. For the qualitative content analysis of the interviews and the comments, a coding scheme (see Appendix) was developed and applied to both document types. The categories were developed deductively based on the identified a priori dimensions and determinants of how different actors perceive copyright laws, as well as which demands they have for regulation (Mayring, 2000).

Findings

Financing and copyright-related usage practises

As several uploaders in our sample report that their YouTube activity is not only a hobby, but an important source of income, the upcoming copyright reform has a significant impact on their work and can endanger their livelihoods (P3). One of the major concerns is the impossibility of monetisation of some videos due to copyright-related reasons, that is, a Content ID match. Uploaders report that they are often disadvantaged due to mistakes in the automatic system, in contrast to rights holders, who can benefit from the uploaded work (P3, Z6). Since potential mistakes in the automatic recognition of copyright infringements lead to rejected monetisation, and hence financing issues, uploaders consider the rights holders and the video platform to be beneficiaries within the partner program.

In regard to uploaders’ usage practises, several YouTubers produce appropriative content (“let’s play” videos, reaction videos, reviews, etc.), which might be controversial in a copyright-related context. Some of them explain that their work is based on the reuse of existing content; therefore the use of it is unavoidable (J2). “Let’s play” videos are a special case. Since video games usually provide music and images, uploaders producing these videos do not need to search for further existing content to reuse. Moreover, these videos are often licensed, or the use of the work is tolerated by the rights holder for marketing reasons (Beyvers & Beyvers, 2015, p. 797). Nevertheless, some gamers share negative copyright-related experiences:

(…) background music in video games is the only thing that’s included, and that sometimes causes problems in a way. I have a project that involves Lego Lord of the Rings, and it plays the official music by Howard Shore, and of course, it’s copyrighted. And that’s why I couldn’t monetise such videos, (...) because I don’t have the rights of the music. (U9)

This means that even though the uploader had a license to publish content, including parts from the video game, they either erred about the scope of the license, which did not include the music played in the game, or Content ID was not able to recognise that the game and the music fall under the same license. In the last case, the rights holders of the music could benefit twice – on the one hand through the already concluded licensing agreement and on the other hand from monetisation. In contrast, content creators are at a disadvantage; although they have contractual permission to use the works, they do not participate in the advertising revenue generated by the video.

Knowledge, experience and evaluation of previous copyright regulations

General considerations regarding copyright regulations

Uploaders who produce and upload appropriative content stated that they are particularly aware of potential copyright issues because they want to monetise their videos. Some of them admitted that they started to think about copyright when they became part of YouTube’s partner program – thus, their channel and their activity on the platform became more professionalised (L3). Similarly, some of the uploaders stated that they were concerned about copyright because they want to be on the “safe side” and do not want to get any strikes against them.3 The third reason for copyright considerations given by the interviewed content creators is of a rather value-driven nature. Some of the YouTubers put themselves in the position of the rights holders and admitted that they would also not like their content to be “stolen” (K8, J2):

I don’t want anyone to just take my videos and then upload them and make money with them. As an author, I would find that unacceptable, and I think it is perfectly okay if the videos are taken down and, in a worst-case scenario, such channels disappear from the platform. (J2)

In addition, some YouTubers emphasised that reusing already existing content and creating new works from it is part of the culture on the video platform. One participant who has run his channel for eight years, uploading mostly reaction videos, stated:

I think that’s the beauty of YouTube, even with reaction videos (...) you can probably ask a lot of YouTubers and most likely 90 percent will have absolutely no problem with people reacting to their video and the others also making a profit with the video because it’s simply a matter of giving and taking. (L3)

This so-called “give and take culture” was also mentioned by other uploaders who claimed to reuse existing content, such as memes and music, but as a reward they tag the initial authors and thus help them become more visible (K8). Nevertheless, this noninstitutionalised culture can lead to copyright conflicts. In this sense, several content creators consider YouTube’s regulatory initiatives too strict and feel that they restrict their creativity (M3, N9, L3):

YouTube used to be called “YouTube Broadcast Yourself” and I think that’s the principle behind which YouTube should still stand … Everyone can do what they want. (L3)

In general, everyone we interviewed agreed that copyright is important due to the aspect of fairness:

I think that’s important, especially on the Internet. For example, I do not work in the music industry, but I know some people there. And these people not only put a lot of energy and passion into their music but sometimes also a lot of money and should be paid fairly. (O6)

Even though uploaders consider copyright regulations to be generally important, they point out that these should not be too strict or unfair, as they otherwise might endanger the participatory culture, which is a core characteristic of YouTube and significantly determines its appeal.

Experience with and evaluation of content ID

All uploaders had some experience with previous copyright regulations to share since all of them had at some point been notified that Content ID matched their uploaded content. In this vein, the question about their experience with previous copyright regulation was directly associated with YouTube’s private regulation due to the absence of adequate state regulation.

Some uploaders stated that for them it was never quite clear how exactly Content ID works, or how their video is verified. Similar to the results in Kaye and Gray’s study (2021), a lack of detailed knowledge leads to speculations about the efficacy of YouTube’s automatic management system. Based on her/his observations, one of the participants claimed that uploaders could theoretically trick the system if they turned the music down very low or talked very loudly during the video (U9). Another participant believed that the Content ID system varies in terms of its effectiveness depending on the content type:

I think when it comes to music and sound, these algorithms or whatever is behind these systems are very efficient. (...) I think it’s much more difficult with visual material because you can choose a different frame, you can somehow cover or change it, and I think it’s much more difficult to find out whether you’re using someone else’s visual material (…). (S6)

Similarly, some of the content creators considered the susceptibility to errors as the automatic system’s most significant risk, which often favours the interests of big production companies and publishers, leading to overblocking and restrictions on creativity (J2, H6). One of the participants added that a possible solution to this problem could be the manual verification of content by humans. When individual case decisions are necessary, human intervention is especially recommended (P3).

Nevertheless, other uploaders emphasised the opportunities that Content ID offers, claiming that the system helps them to identify inconsistencies in their content (K8, L3). However, uploading an entire film, self-recorded in the cinema, is perceived as unacceptable and unfair to the rights holders. Hence, in such cases, it is particularly important that systems such as Content ID check the videos for copyright infringements (N9).

The enforcement of copyright regulation, and thus the concrete implementation of the Content ID system by YouTube, evokes criticism. Some of the content creators consider themselves underprivileged within the triangular relationship, claiming that there is a power imbalance between the uploaders and the bigger companies in their role as rights holders, and in case of doubt, the uploaders are the disadvantaged side of the conflict (P3). Many YouTubers identified this issue when they appealed against a Content ID match. They perceived this process as nontransparent (H6, K8, P3, R5, Z6) but also saw themselves as structurally inferior to the rights holders in this context (P3, L3, J2, R5).

All uploaders have diverse experiences with YouTube’s Content ID. The evaluations of the system are ambivalent. It is striking that – even though all interviewees had already come into contact with Content ID – only few details were known about how the system works. This lack of detailed knowledge leads to the perception that the system is very opaque and sometimes unfair.

Knowledge and evaluation of the copyright reform

Even though they had heard of the protests against Article 13 (later Article 17) of the DSM Directive, or even participated in them, the majority of the interviewed uploaders did not feel well-informed regarding the transposition of the directive and the existence of the new regulatory mechanism. Nevertheless, all interviewees had at least some basic knowledge regarding the planned regulatory measures.

Perception and evaluation of upload filters

The strongly organised players argue that in contradiction to the promise of the German government4 to transpose the Directive without implementing upload filters, the UrhDaG does not prevent them (F4). Furthermore, they are concerned that the upload filters will lead to an overblocking and thereby limit freedom of speech (C1). They are also afraid that the implementation of upload filters will constitute entry barriers for new platforms (G4). In this context, some of the strongly organised uploaders focused their comments on the technical capabilities of the platforms, and expressed doubts that platforms could filter all the content uploaded. In addition, some comments emphasised that live streams are particularly jeopardised by this filtering process (H4).

The moderately and weakly organised uploaders who were interviewed do not categorically deny the necessity of upload filters to enforce copyright online. While they recognise the rights holders’ intention to protect their copyrights (L3, Q3, R5), they see problems in the “grey areas” (J8) of copyright-related decisions and doubt the technical maturity of upload filters to make correct decisions in cases where the classification of content as copyright infringement is unclear (L3, M3, T6). Some uploaders see the danger that, once implemented, upload filters can go beyond copyright and stop uploads for other reasons (O8).

Asked how the implementation and enforcement of copyright through upload filters could be optimised, two central demands are identified based on the interviews. On the one hand, there is a call for more human engagement in the filtering process, at least in the form of a quick and simple human decision-based appeal procedure in cases of content blocked by upload filters (X5, K8, H6). On the other hand, the second demand calls for YouTube to enter into more license agreements with the rights holders (R5, O6) to make the upload filters redundant, or at least to limit their area of application. Nevertheless, uploaders understand that it is not possible to conclude license agreements with all rights holders (M3).

Perception and evaluation of the de minimis limit

While the introduction of the de minimis limit is partly met with understanding and recognised as a compromise to combine the interests of uploaders and rights holders (O6, S6, T6), this regulatory instrument has been partly criticised. For instance, one participant believes that “it complicates the work, (...) if you have to look out for such details (I8)” to not overstep the 15-second boundary. The chosen limits of: 15 seconds per audio or video clip, 160 characters per text and 125 kilobytes per picture are considered arbitrary, that is, not evidence-based and unsuitable for practise (J2, P3). One participant, for example, stated that for him this is “the embodiment of legislation that misses the point” (P3).

Furthermore, the limits are considered too short to be of significant use for the uploaders (L3, N9). For example, the unit of the kilobyte for pictures, is seen as inexpedient for determining whether a picture is a de minimis use, since the file size of a picture is dependent on the resolution (J2, M3). This point of critique is also shared by the strongly organised uploaders in their comments (I4, F4). However, they disagree whether the de minimis limit is a burden for the rights holders. While one side argues that those simple clips can be considered free advertising and marketing for the rights holders (G4), the other side sees it as an economic strain for the rights holders that primarily favours the platforms because they experience greater benefits when there is more content accessible on their platform (M7). Moreover, some of the strongly organised uploaders argued that the de minimis limit goes against the trend that “short, concise content is increasingly consumed” (E1).

Perception and evaluation of the flagging procedure

The strongly organised uploaders were also critical of the flagging procedure. Some of them admitted that there is a risk of abuse (D1, E1, F4). Moreover, they do not believe that an amateur user is capable of correctly assessing whether his or her upload is a copyright infringement. Therefore, they propose limiting the flagging procedure to so-called “trusted flaggers” (E1). Furthermore, the uploaders who submitted their comments on the draft law argued that in cases of dispute, whether an upload was flagged correctly, the platforms have the power of temporary judgement, and consequently the final decision about the content. This leads to private companies being responsible for the protection of human rights (H4).

In addition, one of the comments drew attention to the fact that the flagging procedure is not applicable in certain cases of overclaiming by a rights holder – for example, when a rights holder claims works of public domain that are not (any longer) protected by copyright. In this case, the uploader cannot flag the content as “not protected by copyright”, which creates gaps in the protection at the expense of the uploaders (F4). Weakly or moderately organised uploaders perceived the flagging procedure as complicated, but in general they evaluated it positively. Several YouTubers argued that it is a first step to limit the application of upload filters and to prevent overblocking. In this context, some of them emphasised the importance of the flagging procedure, since the video is immediately uploaded and they do not have to first unlock a blocked video in a complicated process (O6, L3, R5, Y8, J2). Furthermore, some of the interviewed uploaders compared the flagging procedure with the de minimis limit and found flagging more useful (N9), especially in the grey area where a case-by-case decision is necessary (S9).

Nonetheless, it is important to note that a few of the stated demands (in their comments) from the group of strongly organised uploaders have been incorporated into the law. In the comments on the first draft of the law, it was originally argued that the flagging procedure did not include a stay-up obligation in case the rights holder appealed the flagging (F4, G4, M7). This stay-up obligation has been integrated into Section 9 para. 1 of the UrhDaG. It states that the questionable content has to be on public display until the end of the appeal procedure. In addition, some of the uploaders mentioned the risk that platforms would tend to rule in favour of the rights holders to avoid any liability. This concern has been weakened by the inclusion of a release of liability for decisions made in the appeal procedure by the platforms in Section 12 para. 2 of the UrhDaG. Furthermore, the risk for abuse of the flagging procedure by the uploaders has been diminished by the introduction of the so-called red button in Section 14 para. 4 of the UrhDaG.

Perception and evaluation of the red button

The red button was introduced as a reaction to the criticism by the rights holders and strongly organised uploaders regarding the risk of abuse carried by the flagging procedure in the last version of the draft. Several weakly and moderately organised uploaders evaluated the regulatory instrument positively since they considered it fair towards rights holders who can therefore better protect their copyright (J2, K8, Q3, R5, S6, W9, V3). Moreover, the uploaders perceived it positively, as it can prevent so-called “spoilers” that they do not want to be exposed to (R5). However, some of the interviewed content producers consider the red button a valued addition to the law, as long as its area of application is limited. In this vein, they argued that the term “especially trustworthy” is not adequately defined; thus, this regulatory measure opens up the application area too wide and carries the risk of abuse (L3, O6, P3, Q3, T6). Additionally, one of the participants argued that the evaluation of this regulatory instrument depends on “in which cases this will be applied and how often it will be used” (K8).

In conclusion, the different mechanisms are evaluated differently by the uploaders. While some are perceived as helpful for maintaining creative content production (e.g., the de minimis limit), others are seen more critically (e.g. upload filters) as they strengthen the position of the platforms.

Demands on copyright regulation

A central demand of copyright regulation, voiced by the interviewees, is the call for more transparency regarding the legislative process, where “the impression arises that the rights holders, who perhaps have the most resources behind them, are actually in a better negotiating position from the outset” (H6), but also concerning the mode of operation of some regulatory mechanisms, as well as YouTube’s communications. Uploaders want to know exactly which kind of content is harmless in terms of copyright infringements and which is not (K8, P3, R5, S6, W9, Z6). Likewise, several YouTubers expressed a demand for the introduction of a fair use clause into German, or rather, European copyright law, similar to the concept used in the United States. A fair use clause would lead to more transparency and would make producing and uploading videos less complicated (N9). Whether the uploaders did not realise that Section 24 of the UrhG (old version) already had a narrower general clause in the form of “free usage” (Peifer, 2014, p. 89), or whether they did not consider this to be broad enough, unfortunately, did not become clear in the interviews.5

Furthermore, uploaders wished to have the option to negotiate more easily with the rights holders within YouTube’s infrastructure – for example, to conclude individual license agreements (J2, R5): “If I want to use a song in one of the videos, I would like to have a direct dialogue with that person” (R5).

At the same time, some uploaders believe that platforms like YouTube should take more responsibility, and that their interests should be favoured too, similar to the interests of the rights holders. Because platforms provide the digital space where content can be used, they should be responsible for all security aspects on their website, making sure that no rights are infringed upon (H6, I8, N9). Nevertheless, one of the participants highlighted one of the risks of making platforms more accountable for the content:

Yes, I think that making the platform completely responsible bears the great risk for me that the platform takes the only logical step and says, “OK. Then we’ll implement such rigorous filters that somehow, a priori, anything that only appears to possibly violate copyrights may not be uploaded because, in the end, we bear the responsibility for it. (P3)

Conclusion

Since the adoption of the DSM Directive has brought European-wide protests against the introduction of upload filters, in this study we aimed to determine how differently organised uploaders perceive copyright regulation at a time when the reform is being implemented and what further demands on regulation they have. Through the combination of qualitative interviews and a content analysis of comments submitted on the draft law, we can provide a comprehensive picture of uploaders’ perceptions and demands of copyright regulation. Moreover, by interviewing moderately and weakly organised uploaders who did not have the legal expertise to take part in the consultation process on the draft law, we fill a gap in the public discussion on the German Act by giving voice to the people who are an active part of the YouTube community.

Our findings show that the majority of uploaders recognise copyright regulations as important and necessary. Copyright conflicts are an issue on YouTube, and therefore uploaders are concerned about it when they produce and upload their content. In contrast, uploaders also emphasise that copyright regulations must be designed sensitively so that the participatory culture and artistic creativity, which are core characteristics and form the central appeal of online video platforms, are not restricted.

However, the boundaries of the “triangular relationship” seem to blur. Uploaders also take the role of rights holders regarding their creations and therefore want to protect their content on the platform. In this vein, similar to the idea of a privacy paradox (Barnes, 2006), a so-called “copyright paradox” might emerge on video platforms. This means that uploaders want to use as much content as possible, but at the same time, they do not want their content, for which they are the rights holders, to be used for free.

When asked about their experience with regulations, uploaders directly referred to YouTube’s private regulations and did not mention any of the regulatory measures in force. According to our findings, private measures to protect copyright, introduced and implemented by private actors, such as platforms, are perceived as less problematic than the upcoming legislative measures in the wake of the copyright reform, even if they are based on technically identical filter mechanisms.6 The de facto implementation by legislators that obliges the platforms to use upload filters is considered restrictive toward freedom of expression, while filtering measures introduced by private actors on their own accord are perceived as a condition to participate on the platform and benefit from the YouTube partner program. Surprisingly, uploaders do not see a problem with the fact that in the latter case the power to decide about the application of copyright regulations lies solely in the hands of major platform companies (see also Schillmöller et al., 2022). Since upload filters bear the risk of overblocking due to restrictively programmed filters, uploaders demand more investment in automatic filtering systems and/or human intervention, which can help to avoid technical errors. Also, some of the uploaders seem to think that a “fair use” clause – oriented on the US model – could lead to more transparency and would increase the efficiency of the uploading process. However, this conclusion fails to recognise that in this case the interpretation and definition of the term “fair use” would still be completely unclear, and would probably have to be shaped by case law in a lengthy process. In the meantime, it would not offer sufficient protection. Moreover, it seems questionable whether the upload filters – in their current state of development – would be able to recognise whether a case falls under the “fair use” clause. It is unclear if a “fair use” clause will really lead to less overblocking, or if the uncertainty that accompanies such a blanket clause will even increase the danger of overblocking.

Consistent with previous literature, YouTube’s internal appeal procedure and its impact on the perception and enforcement of users’ rights seem to need further investigation.The procedure is considered unfair and could produce chilling effects (Hui, 2021; Kaye & Gray, 2021; Soha & McDowell, 2016), particularly because it not only influences the financial and professional opportunities of users, but also because it can have an indirect impact on freedom of expression. To resolve this, regulation needs to pay more attention to safeguarding procedural rules to ensure a fair trial.

Regarding the implementation of the DSM Directive into German law and the concrete measures that follow from it, it seems important to observe to what extent overblocking can be prevented through the regulatory mechanisms in the UrhDaG. It also remains debatable whether the red button criticised by users is as vulnerable to abuse as feared. To prevent abuse, one option would be to define the vague term “particularly trustworthy rights holders” more precisely in the law. Finally, taking into consideration the uploaders’ needs, more participation and transparency, both in the legislative process and in concrete copyright regulation, would be desirable. A first approach could be to classify the most common usage practises according to their lawfulness and for them to be communicated clearly, either by the platforms themselves or by administrative bodies. In addition, more information about the algorithms used to filter uploaded content could increase their transparency. However, the uploaders are also obliged to inform themselves about existing regulations, their applications and planned reforms. Furthermore, direct contact and dialogue with rights holders would eventually help to restore balance along the conflict lines.

Of course, our study has limitations. Against the background that active uploaders who took part in the protests against the DSM Directive were silent in the public discourse, our analysis solely took their perspective. Since the rights holders and the platforms are considered powerful and influential during the lobbying process (Beisel, 2019), their stakeholders should also be included in the research to clarify and contextualise the political decision-making process at the EU level. Furthermore, the study focused on the German example, shedding light on the perception and demands of German uploaders. Due to the specific topic and the nationally binding nature of law, focusing on only one country was necessary to provide an in-depth analysis of the different regulatory measures and their implementation in national law. However, cross-national comparative studies could widen the evaluative perspective and the spectrum of copyright regulation options.

In sum, the present findings confirm the need to narrowly observe and evaluate the implementation of copyright reforms. Nevertheless, it is crucial to adapt copyright regulation to the current digital dynamics while protecting both democratic values and the interests of different conflicting parties. Future interdisciplinary research could contribute to identifying potential challenges regarding the concrete implementation of reform at an early stage, developing evidence-based regulatory measures or adapting the current ones.