11.1 Introduction
Disinformation, an old phenomenon, has experienced a resurgence with new characteristics and significant disruptive capabilities in today’s information landscape, greatly influenced by technology and increasingly dominated by a small number of global digital intermediaries. The potential of the large-scale and easy spread of disinformation online is considered a risk for freedom of expression, in all its aspects and corollaries, and ultimately to democracy,Footnote 1 as well as to humanity.Footnote 2 If combined with other contextual elements, such as the opacity of algorithm-driven online news intermediaries, the dominance of major online platforms, the polarisation of public discourse, and the declining viability of journalism and traditional news outlets, disinformation emerges as an increasingly significant disruptor, threatening fundamental rights, information diversity, electoral integrity, and democracy itself. The picture is even gloomier if we consider how the further spread of disinformation seems one of the main challenges posed by generative artificial intelligence whose development and applications are growing exponentially. From a policy perspective, it is therefore deemed necessary to address this risk to mitigate its impact on public opinion and fundamental rights, while also ensuring the protection of the fundamental right to freedom of expression, which can be compromised by inadequate regulatory interventions.
Addressing disinformation is a challenging task due to the difficulty of defining and identifying disinformation in a clear way, the sheer volume of content online and the speed of its dissemination. To a certain extent, this context makes it inevitable to rely on platforms in addressing such content or behaviour. Furthermore, no one knows what happens behind the walls of large platforms, as access to data for researchers and regulators is limited and much of the information falls under trade secrets. This results in information asymmetry between regulators, especially at the levels of individuals and smaller states, and the online platforms that have positioned themselves as key actors in the online information sphere. When a government cannot know as much about the industry as the industry knows about itself, and where no single actor has all the information needed to understand the problem and all the instruments needed to solve it, various types of ‘decentred regulation’ emerge as solutions.Footnote 3 This ‘decentred understanding of regulation’, as Julia Black has articulated it, moves from state-centred regulation to various forms of self-regulation.Footnote 4 In addition to this, even when forms of more traditional regulation emerge, they need to rely on private actors – in this case online platforms – for enforcement. While such a regulatory shift may be the only means to address the complexity of contemporary challenges, it often itself encounters challenges related to the sufficiency of oversight and the accountability of platforms, especially as these platforms become both the regulated and the regulator and as they make major decisions with impacts on fundamental rights.Footnote 5
In this chapter, we explore the complex interplay between the need to combat disinformation and the importance of safeguarding fundamental rights, particularly media freedom and media pluralism, as essential aspects of freedom of expression and the role of the news media as ‘the fourth estate’. We do this by examining the key European Union instruments in this area, namely the Digital Services Act, the Strengthened Code of Practice on Disinformation, and the European Media Freedom Act. These instruments connect on the matter of curbing disinformation and protecting media freedom, as a key value upon which the EU and other democratic systems are built. The chapter begins by highlighting the key challenges of today’s information environment, such as disinformation and the dominant role of online platforms, which significantly affect media pluralism and citizens’ access to journalistic content. It then outlines the evolution of the EU’s policies and regulations in the area of information integrity. In the final sections, the chapter examines the key elements of the EU’s approach to tackling disinformation, particularly in relation to media pluralism, and the challenges associated with this approach.
11.2 The Contemporary Information Environment: Platforms, Media, and Disinformation
Citizens’ habits around consuming news and information that are critical to forming political opinions have been shifting for some time, in some countries and between younger demographics increasingly favouring digital intermediaries over direct access to news media.Footnote 6 Leading online platforms have stabilised their role as key intermediaries between media outlets, their audiences, and advertisers. Through their content policies, platforms control the visibility of news and media content, strongly influencing its potential for monetisation by advertisers. By distributing the media content, platforms economically benefit from it and at the same time compete with media content providers in the online advertising market. They hold a significant advantage in this competition due to the vast amounts of data they collect, enabling them to better understand and target users and making their services as well as their overall reach more appealing to advertisers. As shown by the European Media Industry Outlook report, platforms such as Facebook and Google capture major shares of digital advertising revenues, while the news media lost more than 20 per cent of their revenues between 2016 and 2021.Footnote 7 Such market dominance and the gatekeeping role of platforms, as corporate actors with limited liability, threaten not only the economic sustainability of the media but also the democratic information environment and media freedom as constitutional values.
The online-platform-driven information environment has given rise to new voices and content creators who may not respect principles of accuracy and verification but who can still significantly influence public opinion. While the diversity of voices facilitated and amplified by online platforms has improved pluralism in the public arena, it has also led to high and globally shared concerns over information overload,Footnote 8 news avoidanceFootnote 9 and information disorder.Footnote 10 Influencers across various platforms are blurring the line between professional and non-professional content, with fewer safeguards to ensure accuracy. This shift contributes to the phenomenon of disinformation: the intentional spread of false or misleading content for economic or political gain. Even when the spread of false or misleading content is unintentional, the effects can still be harmful. As explained by Gershberg and Illing, this is one of democracy’s paradoxes: the more open communication is, the more demagoguery and misinformation follow.Footnote 11 Disinformation, on the other hand, is strategic and coordinated, using technology in harmful ways and exploiting the vulnerabilities of both citizens and the broader information environment.
Online platforms, and in particular social media platforms, have emerged as powerful actors in the information sphere – a new type of entity that cannot be classified as traditional media companies or publishers but also that do not function as neutral technology companies or mere hosting services that could easily fall under the liability rules such as Section 230 of the US Communications Decency Act or the European E-Commerce Directive. In the aftermath of the Cambridge Analytica scandal, concerning the 2016 US presidential elections and the UK’s referendum on leaving the EU in the same year, the potentially disruptive impact of online platforms on democracy and legal responsibility came into the spotlight. Lawmakers began publicly questioning platform leaders – such as Mark Zuckerberg of Facebook (now Meta) – about the legal nature of their services. The responses from platform executives have been confusing and inconsistent:Footnote 12 while claiming that Facebook is a technology company, not a media company,Footnote 13 they have also invoked publisher rights and protections in certain court cases.Footnote 14 Overall, platforms have been relatively more willing to publicly acknowledge responsibility,Footnote 15 but less willing to accept liability for the content and behaviour they mediate and their effects. As detailed by Bertolini et al.:
Responsibility, as typically referred to in the law and policy debate revolving around advanced technologies, is a wider concept than that of liability, and it is often used to denote the moral responsibility of a subject, as defined by the philosophical, sociological or political debate. On the contrary, liability denotes that specific form of legal responsibility that is connected to the violation of a duty that the person held liable was obliged to comply with, or to the infringement of one’s rights.Footnote 16
In the EU framework, the general conditions for platform liability were outlined in the E-Commerce Directive, which provides a liability exemption (Article 14) for service providers primarily acting as hosting services.Footnote 17 The exemption applies if the platform (a) does not have actual knowledge of illegal activity or information, and, in cases involving claims for damages, is not aware of facts or circumstances that make the illegal activity or information apparent; or (b) upon obtaining such knowledge, acts promptly to remove or disable access to the illegal information. Scholars such as Helberger, on the other hand, argue that regulators should view platforms not as mere ‘intermediaries’ but rather as active political actors wielding significant influence over public opinion.Footnote 18
More recent EU regulatory interventions, such as the Digital Services Act, require platforms – in particular those that, because of their size and reach, are very likely exerting an influence on public opinion – to take a more proactive role in combating disinformation and assessing and mitigating other systemic risks to society, democracy, and individuals’ rights and well-being.Footnote 19 This includes any actual or foreseeable negative effect on media freedom and media pluralism deriving from platform use or the very design of their services. Media and journalism play dual roles with respect to disinformation and broader information disorder challenges. They can both contribute to the spread of manipulative narratives through unverified or sensationalised content and serve as a critical tool for combating disinformation by providing accurate, complete information. Trusted, independent media are essential for tackling disinformation, but declining professionalism and economic pressures have often made the media part of the problem. This is why our analysis of EU policy and of the regulatory framework for addressing disinformation and safeguarding the integrity of the information environment considers both the platforms’ duty of care for information integrity and fundamental rights and the role of the media in key anti-disinformation measures.
11.3 Evolution of the EU Regulatory Framework for Information Integrity: From Self-Regulation to Co-regulation and Regulation
Threats to information integrity have become a matter of global concern because access to information is one of the fundamental rights on which many other rights and obligations of individuals in a democratic society are based.Footnote 20 In the EU, the exposure of citizens to disinformation at scale has been recognised as a major challenge for democracies, one that requires a policy response.Footnote 21 However, addressing disinformation is a complex policy task. It first encounters the challenge of disinformation being a multifaceted phenomenon that cannot be easily operationalised and categorised within existing policy frameworks and competencies.Footnote 22 Disinformation can be approached from the angle of national security, election integrity, public health, informed citizenship, and democracy. When addressing disinformation, it is crucial to consider the range of actors involved, each with varying degrees of agency, awareness, and intent. Disinformation spans from illegal activities to legal but harmful speech, often intersecting with the boundaries of freedom of expression. The transnational nature of online disinformation is evident, yet it is deeply influenced by national contexts, legal traditions, and political cultures.
The European Union is at the forefront of policy making to address disinformation, and as such inspires regulatory activities worldwide.Footnote 23 At the core of the EU’s approach in this area is the stated commitment to the protection of fundamental rights, including media freedom and pluralism. This policy area includes various instruments, such as data protectionFootnote 24 and the regulation of political advertising,Footnote 25 as well as development of media literacy skills,Footnote 26 but the strongest emphasis has been on the roles and responsibilities of large online platforms. This is why this analysis focuses on the Code of Practice on Disinformation, the EU’s primary tool for tackling disinformation, which was first adopted in 2018 and revised in 2022.Footnote 27 The 2022 revision includes a commitment to transforming it into a Code of Conduct under the Digital Services Act (DSA) for very large online platforms, signatories of the Code.Footnote 28 This chapter, therefore, analyses relevant provisions of the DSA and its interaction with the European Media Freedom Act (EMFA) concerning the relationship between media, platforms, and disinformation.
11.3.1 Self-Regulatory Framework: (Strengthened) Code of Practice on Disinformation
Self-regulation is often considered more adaptable than formal regulation, and potentially more suitable for addressing evolving challenges, where formal regulation might be overly rigid and intrusive. While there is no unified definition, self-regulation is commonly understood as a regulatory process wherein industry actors establish rules and standards governing their conduct within the industry.Footnote 29 The first policy instrument born under the EU’s approach to tackling disinformation was the 2018 Code of Practice on Disinformation.Footnote 30 It was initiated by the European Commission (government),Footnote 31 and it marked the first instance globally where the industry agreed to self-regulatory standards in response to disinformation. Initially, the Code was signed by Facebook, Google, Twitter, and Mozilla, along with several entities representing the advertising industry. Microsoft and TikTok joined later, and Twitter withdrew after becoming X, illustrating the Code’s fragile nature.
The signatories committed to implementing a set of measures to address the problem of disinformation. These measures revolved around five areas of commitment: (i) scrutiny of ad placements, which focuses on demonetising the dissemination of disinformation; (ii) transparency of political advertising, committing to making political or issue ads clearly labelled and distinguishable and requiring a verification process for those who purchase such ads; (iii) ensuring the integrity of platform services by limiting manipulative behaviours and practices, such as impersonation, malicious deepfakes, and the purchase of fake engagements; (iv) empowering users through media literacy activities, transparency of recommender systems, and user access to tools and information to evaluate the trustworthiness of information sources; (v) empowering the research community by providing and facilitating access to platform data for the study of disinformation and of the effectiveness of measures against it.
Following the initial period of the Code’s implementation, it became clear that its content and implementation framework needed to be revised to allow for appropriate monitoring of its effectiveness and broader impact. The review of the 2018 Code, conducted by the European Regulators Group for Audiovisual Media ServicesFootnote 32 and the European Commission,Footnote 33 highlighted a set of significant shortcomings. These included the lack of clear and consistent definitions of key concepts, starting with the concept of ‘disinformation’, which some platforms approach primarily as content while others observe it as coordinated inauthentic behaviour. Additionally, there were challenges in assessing the activities and their impact due to limited access to functional data from platforms. Key performance indicators (KPIs) had also not been established, making it difficult to measure achievements effectively. To address these issues, an updated version of the Code was introduced in summer 2022, featuring a broader range of signatories, more detailed commitments, and KPIs. This update was designed with the implementation of the Digital Services Act in mind, announcing in the Preamble of the Code of Practice its transformation into the Code of Conduct under the DSA. The Code is intended to serve as a tool for fulfilling the obligation to mitigate the systemic risks posed by disinformation, as outlined in Articles 34 and 35 of the DSA. The 2022 Strengthened Code continues to be a voluntary instrument but it is increasingly co-regulatory, with the European Commission’s involvement in overseeing its implementation.
The primary signatories of the Code continue to be very large online platforms and search engines. However, the 2022 Strengthened Code was expanded to include a broader range of signatories, including civil society groups, fact-checking organisations, and some commercial entities focused on trust and safety in the online environment. News media and journalistic organisations are notably absent.Footnote 34 This questions the effectiveness and wider implications of the instrument, especially as it is integrated withing the DSA as a potential compliance mechanism.
11.3.2 Co-regulation on Disinformation: The Digital Services Act
The Digital Services Act is an EU regulation that came directly into full application across the Union from mid-February 2024. It seeks to enforce consistent rules for intermediary services across the internal market in tackling the spread of illegal content and systemic societal risks such as disinformation. Because disinformation and other forms of manipulation in the public sphere are harmful but not always illegal, this EU Regulation applies a different approach to illegal content as compared to content that is legal but harmful. At the core of such an approach is the need to protect fundamental rights, including freedom of expression of the users of digital services. Among the key concerns and challenges in effectively enforcing the DSA, however, are the complexity of the online environment, the vast volume of content circulating on online platforms, and the challenge of swiftly determining legality or illegality.
Content that is harmful, but not necessarily illegal, as well as a number of systemic risks that represent a challenge to freedom of expression, media pluralism, informed citizenship, and electoral integrity, fall under the duty of care of very large online platforms (VLOPs) and very large online search engines (VLOSEs): those that have more than 45 million users per month in the EU. Article 34 of the DSA outlines the risk assessment obligations for such platforms and search engines, requiring them to identify, analyse, and assess systemic risks arising from the design, operation, or use of their services. These assessments must be conducted at least annually and before deploying functionalities that may critically impact identified risks. The assessments should be specific and proportionate, taking into account the severity and likelihood of systemic risks. The DSA highlights four categories of systemic risks: (1) dissemination of illegal content; (2) negative effects on fundamental rights; (3) negative effects on civic discourse, electoral processes, and public security; and (4) negative effects related to gender-based violence, public health, and minors.
Disinformation is of concern in two out of these four broad categories of systemic risks: it threatens civic discourse and electoral processes, and it undermines fundamental rights such as freedom of expression and access to (accurate) information. Any actual or foreseeable negative effects on media freedom and pluralism are also considered a systemic risk. These negative effects may result from disinformation and manipulative use of platforms services, but also from content moderation policies by large platforms that, in an effort to reduce disinformation, decide to limit the reach of political and media content overall.Footnote 35
In addition to assessing systemic risks, VLOPs and VLOSEs are explicitly required to mitigate them by Article 35 of the DSA. The mitigation measures should be reasonable, proportionate, and effective, considering the impact on fundamental rights. However, this is a completely new area of action for both the platforms and the European Commission, which is supervising and enforcing the DSA for VLOPs and VLOSEs, especially in relation to systemic risk assessment and mitigation. How to operationalise such broad categories of risk, and which standards and benchmarks to use in determining whether a risk exists and if it is systemic, has been largely left to the platforms, at least in the early stages of the DSA’s implementation. Given that platforms are primarily commercial entities and, under the Code of Practice, were unable to agree on a common definition of disinformation, it is reasonable to be concerned about delegating the critical task of assessing systemic risks and fundamental rights to them.
There are certain oversight mechanisms foreseen in relation to risk assessment and mitigation conducted by platforms. Article 37 of the DSA requires VLOPs and VLOSEs to subject themselves and their services to independent audit, at their own expense and at least once a year. However, the methodology for these audits has remained unclear in the first year of the DSA’s implementation. What is known, as required by law, is that audit reports must be shared with the Commission and the Digital Services Coordinator of the platform’s EU establishment. These reports must include the results of the risk assessment and the specific mitigation measures put in place. Additionally, Article 42 mandates that VLOPs and VLOSEs provide transparency reports every six months. Article 40 requires platforms to grant researchers access to data necessary for monitoring and assessing compliance with the DSA, particularly in relation to systemic risk assessment and mitigation.
The implementation of the Code of Conduct on Disinformation must also be measurable and subject to public oversight. General oversight should be conducted by the Commission and a network of competent authorities, with the Digital Service Coordinators (DSCs) serving as the primary authorities designated by Member States. The DSCs collaborate at EU level through the Board for Digital Services and at the national level with other relevant and competent bodies. This is a large network that will take time to fully establish and develop clear processes in an area that is still relatively new to all involved. In cooperation with the DSCs and following public consultations, the Commission may issue guidelines in relation to specific systemic risks and good practice in addressing them. Furthermore, the European Board for Digital Services, in collaboration with the Commission, must annually publish comprehensive reports that identify and assess recurring systemic risks in these platforms and search engines, along with best practices for mitigating these risks, categorised by Member States and the Union as a whole.Footnote 36 Platforms are required to retain supporting documents for their risk assessments for at least three years and to share them with the Commission and the Digital Services Coordinator upon request.Footnote 37
The Commission can also open a formal proceeding or investigation into a potential breach or lack of compliance with DSA obligations, including on risk mitigation; there needs to be justified suspicion, which presupposes a detailed insight into the operation of various platforms that fall under the category of VLOPs and VLOSEs. Platforms are known for their frequent changes of content and policies, which may arise from external pressures including legal requirements, but will seek to align them with their internal economic considerations. Currently, there is a clear information asymmetry between online platforms and most other relevant stakeholders, including the media, regulators, and policymakers, which will be difficult to tackle even with the mechanisms provided by the DSA. Systematic monitoring of platform activities and their impact requires substantial investment and expertise, which many regulators and researchers across the EU may find challenging to acquire. Therefore, as it stands now, the DSA may strengthen platforms’ own supervision of media content, media freedom, and pluralism, especially if strong and systematic public oversight is not fully ensured. The DSA, in the end, together with the Code of Practice on Disinformation, significantly entrusts very large online platforms to define and assess the key concept of European media policy – media freedom and pluralism.
11.3.3 The Other Side of the Coin: The European Media Freedom Act
In spring 2024, a European law that seeks to protect media freedom and pluralism was adopted, which is a groundbreaking achievement in itself given that the EU has limited competences in the area of media policy.Footnote 38, Footnote 39 While the main objective of the European Media Freedom Act (EMFA) is to improve the functioning of the internal media market, it eloquently emphasises the duality of the media business, which is not and cannot be seen as any other business.Footnote 40 In addition to their market dimension, media services play a vital role in safeguarding fundamental rights such as freedom of expression and access to information. When they operate independently and in a pluralistic manner, they foster informed citizenship, which is essential for the functioning of democracy.
The EMFA mostly addresses undue political influences and market conditions for the independent and viable functioning of the media. Some of its key provisions include obligations for media ownership transparency, safeguards to ensure the independence of public service media, nuanced assessment of media market concentrations, and the establishment of the European Board for Media Services to enhance cooperation among national media regulatory authorities. It also establishes an important principle that media content is not just any other content on online platforms and should thus be handled with special care. Article 18 of the EMFA is the only provision at EU level to recognise media and journalism as key democratic institutions that should not be easily restricted by technology companies, which have no editorial responsibility or legal liability for the content they host, unlike the media. As stipulated by Article 18, before a VLOP suspends or restricts the visibility of content from a media service provider (who self-declares as such under a functionality provided by platforms), it must provide a statement of reasons for its decision and allow the media service provider to respond within twenty-four hours. If no response is received and the platform decides to proceed with suspending or restricting visibility, it must promptly inform the media of its decision.
Article 18 of the EMFA contains safeguards to ensure that this special treatment is not misused by actors who do not comply with professional and regulatory standards in one or more EU Member States. Furthermore, the list of media benefiting from this special treatment should be easily publicly available to allow for monitoring of who benefits from this provision. These safeguards were developed as the result of a heated debate around this provision and concerns voiced by experts and stakeholders involved in the area of tackling disinformation.Footnote 41 The concerns primarily revolved around the method of identifying media service providers through self-declaration. There was apprehension that this special treatment, which involves a delay in content moderation, could be exploited by malicious actors, including state-controlled or propagandistic outlets, to spread disinformation, misinformation, and propaganda more effectively. Consequently, such practices could undermine the implementation of the Digital Services Act in assessing and mitigating the risks associated with disinformation.
The final result in the EMFA on Content of media service providers on VLOPs (Article 18) is that the special treatment for media does not apply where providers of very large online platforms suspend the provision of their services or restrict the visibility of media content in compliance with their obligations pursuant to Articles 28 (measures to ensure a high level of privacy, safety, and security of minors), Articles 34 and 35 of the DSA (systemic risk assessment and mitigation), and Article 28b of the Audiovisual Media Services Directive (requiring video-sharing platform providers to protect minors and the general public from harmful and illegal content) or with their obligations relating to illegal content pursuant to Union law. Given the broad definition of systemic risks under the DSA, along with other exceptions, there are very few instances in which the special treatment for media, as outlined in Article 18 of the EMFA, may be applicable. Nevertheless, it remains an important principle that media content is a distinct category on online platforms.Footnote 42 Despite its likely limited impact, this provision in the EMFA upholds media freedom in the environment of online platforms where both the democratic function and the economic sustainability of the media are increasingly under threat due to the growing power of private actors in content moderation and shaping the information sphere.
11.4 Defining, Privileging, and Including Media
The media are crucial actors in the realm of disinformation, capable of both exacerbating the problem when operating with low professional standards, and serving as a key solution when carrying out their democratic task responsibly. It is surprising, therefore, that the media are not part of EU’s key instrument for addressing online disinformation, the Code of Practice on Disinformation. It is unclear whether there was no direct or compelling invitation extended to the media to participate in the Code, or if the media did not perceive themselves as part of this initiative. In any case, a policy instrument promoting self-regulation and multi-stakeholder involvement must be evaluated based on which stakeholders are included, as this significantly impacts the quality of policy solutions and may have broader consequences. By tackling disinformation, the Code implements certain measures that can impact the media and their content in the environment of online platforms, which have become an important gateway to news.
The measures and impact of the entire 2022 Code are important for the media, just as the media are important for the fight against disinformation in both online and offline environments, which increasingly blend. There are several commitments within the Code that can specifically affect the media, their reach on online platforms, and their potential for monetisation. These are primarily ‘indicators of trustworthiness’Footnote 43 and prominence granted to ‘authoritative sources’.Footnote 44 Relevant signatories commit to provide options to integrate, when commercially viable, ‘indicators of trustworthiness’ to guide ad buyers and demonetise disinformation. Such indicators or signals of ‘trustworthiness’ are also seen as empowering users by enhancing the discoverability of reliable content.
According to the Code, users may be given the option to incorporate signals about the ‘trustworthiness’ of media sources into their recommender systems. There are already companies among the Code signatories that provide ‘trustworthiness’ indicators as one of their business offerings (for example, NewsGuard). They were asked to commit to ensuring that information sources undergo reviews transparently, apolitically, without bias, and independently, with disclosed criteria applied equally to all sources to allow independent audits by regulatory authorities. They also agree to implement correction mechanisms and respect the right of publishers to be heard before applying indicators. These measures and commitments of the Code were established without direct involvement from the media, who are not among the Code signatories. This raises questions about the practical implementation of the measures and the criteria for ranking trustworthy sources in feeds or search results, which remain unclear from the Code. Moreover, it is not clear how the ranking policies will apply to all the sources indexed as trustworthy or how they will balance local, national, and transnational media. Given that ranking policies are closely linked with the media’s reach on online platforms, these questions are pertinent for both media economics and the media’s social role.Footnote 45
Another commitment in the Code relevant to the media focuses on promoting ‘authoritative sources on topics of public and societal interest or during crisis situations’ through features such as information panels, banners, pop-ups, maps, and trustworthiness indicators.Footnote 46 However, it remains unclear who will set the criteria for determining which sources are considered ‘authoritative’ and how decisions will be made about which topics are of particular public interest or crisis relevance across the diverse countries and regions where the platform signatories operate. If these decisions are to be made by the signatories, which include platforms, civil society groups, fact-checking organisations, and commercial entities – but notably not the media – this raises concerns about the exclusion of media from this framework.
By encouraging the adoption of codes of conduct, the DSA provides a tool to convert the Code of Practice into a co-regulation mechanism of compliance with systemic risk assessment and mitigation obligations for the VLOPs and VLOSEs. In this way, the current composition and scope of the Code make it a leading approach in platform regulation on disinformation, interplaying with fundamental rights and media freedom in the platform environment. The way in which VLOPs and VLOSEs assess and mitigate the systemic risks under the DSA affects the scope of implementation of Article 18 of the EMFA, a special treatment for the media service providers in content moderation by VLOPs. This is a complex sequence of actions and responsibilities that shape the conditions for freedom of expression and media freedom through different stages. As Helberger warns, ‘social media increasingly determines the conditions by which the legacy media functions’, and due to the shifting power conditions ‘the fourth estate risks increasingly being unable to act independently from platforms’.Footnote 47 This is why it may be problematic when national or supranational policy and legal instruments ‘formalize the role of social media platforms as the governors of much of the speech that is being shared online’.Footnote 48 Once this role is delegated, it is of the utmost importance that sound independent monitoring is provided, to at least partially restore accountability and enhance trust in a mechanism that otherwise could be perceived as opaque and arbitrary.
11.5 The Challenge of Public Oversight
Information asymmetry between online platforms (as information holders) and researchers, regulators, and journalists (as information seekers) is a bottleneck in the attempts to understand the problem of disinformation and the role that online platforms play in it. As noted by Leerssen, ‘information asymmetry represents the very foundation of economic and societal power’ that platforms hold.Footnote 49 The limited access to platform data and the highly specialised expertise required to assess it make it difficult for regulators, researchers, and civil society actors to effectively understand and monitor algorithmic processes and their impacts across the EU. This creates significant challenges in ensuring transparency – a key principle of governance – especially when the assessment of fundamental rights is delegated to private actors. It is also a stumbling block for the development of evidence-based policies and for any effective monitoring of a policy implemented in this area.Footnote 50 This information asymmetry is recognised as a general problem both across the EU and at the global level, but it also contains a dimension of regional asymmetry in data access and oversight potential: some states, public authorities, and researchers obtain better access due to their political power, legal competencies, and research capacities, while others are lagging behind.Footnote 51
Evaluating policy effectiveness and its impact, including unwanted effects, is a crucial aspect of policy-making. The first iteration of the Code of Practice was criticised for having neither ‘clear and meaningful commitments’ nor ‘measurable objectives’.Footnote 52 This later appeared as a key obstacle in assessing the effectiveness and impact of the Code,Footnote 53 and ultimately also its purpose if there is no possibility of oversight. To address this deficiency, the updated 2022 version of the Code entails a set of key performance indicators, organised in two levels: service-level and structural indicators. Service-level indicators are tied to the specific measures adopted under the Code. More specifically, each measure agreed under the Code is accompanied by one or more service-level indicators that define which information or data is to be provided biannually by relevant signatories in order to monitor the implementation and effectiveness of the measure. Structural indicators, on the other hand, are envisioned as a more objective assessment of the disinformation phenomenon and the effectiveness of the Code in suppressing it across time, both in individual EU Member States and in the EU as a whole. These indicators are intended to provide external, independent assessments rather than rely on platforms’ self-reporting. However, as of the writing of this chapter – more than two years after the adoption of the 2022 Code – such indicators have yet to be fully established.Footnote 54
Article 40 of the Digital Services Act now provides a framework for the transparency and access to data that are necessary to monitor and assess compliance with the Regulation by very large online platforms and search engines. Providers of VLOPs and VLOSEs must grant access to data, upon a reasoned request by the Digital Services Coordinator of establishment or the Commission, for the purpose of monitoring and assessing compliance with the Regulation. Such access must be provided in a manner that respects the rights and interests of the platform providers and service recipients, including data protection and confidentiality.Footnote 55 The Digital Service Coordinator or the Commission may also request platforms to explain the design, the logic, the functioning, and the testing of their algorithmic systems, including their recommender systems. Although this is a significant step forward in terms of data access and oversight potential, without effective methods of validating the submitted data and information it still relies significantly on good-faith behaviour by platforms.Footnote 56
What could serve as a validation mechanism to a certain extent would be allowing access to data for diverse and widely distributed researchers and research groups. Access to platform data foreseen under the DSA extends also to ‘vetted researchers’ (including civil society organisations that conduct scientific research in the public interest) who meet specific conditions and conduct research related to systemic risks and the assessment of the adequacy, efficiency, and impacts of risk mitigation measures. While the technical and security considerations of this access should be defined by delegated acts, it is nevertheless a step towards more information symmetry between online platforms and other relevant stakeholders. As such, it is a key precondition for effective monitoring of the online information environment and the effectiveness of policy designs and measures in this area, including against any unwanted consequences that they may provoke.
At the moment, it remains unclear whether and how a dimension of regional asymmetry in data access will be tackled by the DSA and delegated acts.Footnote 57 This is especially important as certain countries in Europe, as shown by the Media Pluralism Monitor,Footnote 58 exhibit relatively high structural risks to media freedom and pluralism – risks that may be exacerbated online and in the realm of online platforms. Very often in such countries researchers have few resources or capabilities to access platform data and to use it for research that can inform and assess policies. Considering the importance of building policies based on evidence, and having in mind the diversity of political and media systems in EU Member States, the data access provided should be inclusive, and local capacities should be strengthened to benefit from such access. Functional data access is the key dimension of transparency of platform operations and is the precondition for any meaningful assessment of the impacts that platforms’ policies, or measures that they implement as a result of EU policies, may have on information environments.
11.6 Conclusions
Tackling online disinformation requires a sophisticated and nuanced approach, recognising both potential risks to democracy and the importance of safeguarding freedom of expression and other fundamental rights. The European Union’s regulatory framework, particularly the Digital Services Act and the European Media Freedom Act, represents a first, sophisticated, and comprehensive effort to manage these challenges by regulating online platforms, on the one hand, and aiming at enhancing media freedom and pluralism on the other. The DSA’s provisions on risk assessments and mitigation for VLOPs and VLOSEs aim to protect the public from disinformation while respecting freedom of expression. The EMFA, in dealing with the relationship between media and VLOPs, specifically aims to uphold media freedom by ensuring that media content is treated with special care in content moderation processes. However, the effectiveness and the interplay of these regulations depend significantly on their reasonable implementation, on ways to assess their effectiveness adequately, and on ways to avoid arbitrary behaviours, practices, and definitions (of media services, of media pluralism) by VLOPSs and VLOSEs. In short, it is of the utmost importance to avoid the implementation of this corpus of regulations ending in a hidden delegation of power to the platforms, leading to the paradox that the law itself that aimed at making platforms more accountable instead strengthens their role and power. Challenges remain, such as ensuring sufficient oversight and accountability of platforms, avoiding the outcome that the implementation of Articles 34 and 35 DSA are simply elements of a shallow compliance, and addressing the complexity introduced by the rapid development of generative artificial intelligence, which complicates the disinformation landscape further. Overall, it is essential that a balanced approach is adopted in the implementation: this approach should involve multiple stakeholders, including platforms, the media, and civil society. As the landscape evolves, ongoing assessment, adaptation of policies, and international cooperation will be crucial to effectively counter disinformation without compromising the essential values of an open society.
Open access