No innocents: Platforms, politics, and media struggling with digital governance

In retrospect, the communication world was so different in February 2020, when scholarly members of the Euromedia Research Group applied to become a Jean Monnet Network, focusing on media and platform policy (EuromediApp). Shortly after sending off the application, Covid-19 conquered the planet and jeopardized the main objective of networks, namely, to strengthen ties between network nodes. When the three-year network started operating in October 2020, it immediately became clear that dominant features of the pandemic would be fake news and harmful content online. Additionally, it was evident that digital platforms would play an even more central role in opinion-shaping during lockdowns than they had before. During the following three years, it turned out that the concept of the EuromediApp network was smart. Focusing on digital platforms, their relations to mass communication, and their performance regarding democracy and human rights allowed the network to organize cutting-edge workshops and conferences. For these events, it invited scholars to contribute scientific state-of-the-art texts and presentations on this fast-moving topic. This special issue of Communications serves to consolidate the learnings from that journey, timely addressing burning issues in digital platform governance. It explores questions such as how to limit hate speech and other harmful content online, how to hold digital platforms accountable for publishing it, how to accommodate automated decision-making (a.k.a. artificial intelligence), and how to economically balance platform profits achieved at the expense of mass media. Several attempts have been made over the last years to allow digital platform communication to thrive within the boundaries of the wider policy concept of

public interest. Obviously, unprecedented market power wielded by dominant platforms, together with massive waves of harmful and fake content flooding deeply divided societies, called for interventions by states and governments that were fully engaged in dealing with the Covid-19 pandemic.
The five articles in this special issue address platform performance and platform regulation from various angles. No definite and final answers are given by the authors, but they contribute to understanding better the complexity of this vast new field of communication policy.

Combating disinformation using crisis communication: An analysis of Meta's newsroom stories
Michaël Opgenhaffen explains how Meta communicates its responsibility for dealing with disinformation from 2016 to 2022. His approach includes the recent wave of disinformation discourses related to the Covid-19 pandemic and the war in Ukraine. Moreover, he analyses Meta's communication using the theoretical lenses of crisis communication and impression management theories. The disinformation-crisis discourse referred to above puts platforms at the center of the problem, meaning that platforms' reputations are at stake. This crisis impacts platforms' stakeholders such as employees, shareholders, users, advertisers, political actors, and news media, requiring a reaction from organizations that Opgenhaffen interprets as crisis communication.
The article addresses news stories posted by Meta on its newsroom website, which deal with the issue of disinformation. Opgenhaffen concludes that Meta's main message has been consistent throughout these six years: Meta is not responsible for the disinformation crisis but is working on the issue. It partners with fact-checkers, news organizations, and scientists (especially in the context of Covid-19) to correctly inform users, and it develops technology (automated decision-making) as a means of scaling its efforts, often stressing the combination of human and technology as the best strategy to achieve those goals. To reinforce its commitment, Meta constantly mentions its own internal policy. Remarkably, with more weight since the pandemic, Meta emphasizes its efforts to provide good information with partnerships, automated tools, and updates in the community standards, shifting the focus away from the negative presence of disinformation.
The message here is clear, reproducing the playbook of neoliberal governance: no need for stronger public regulation regarding content, as corporate stakeholders themselves exert meaningful self-regulation and are better equipped to tackle the problem without infringing freedom of expression. This contribution adds up to an important understudied dimension of the problem of disinformation and public policy: the strategies the platforms deploy to position themselves and reduce risks for their organizations, especially the risk of stronger public regulation.

Promoting responsible AI: A European perspective on the governance of artificial intelligence in media and journalism
Colin Porlezza's text delves into the ongoing debates within the European Union and the Council of Europe regarding the regulation of artificial intelligence (AI) in news media and journalism. The author highlights the prevailing tendency of news media to present AI in a positive light, while also acknowledging the emergence of a critical discourse. The text critically examines 24 EU policy documents that address the topic of AI. These documents frequently refer to digital intermediaries such as search and social media but largely overlook traditional news media, which are increasingly adopting automated decision-making processes and play an even more crucial role in providing information to EU citizens.
His analysis found that, when news media are mentioned it is typically within the context of associated risks, educational aspects, social responsibility, and the role of media as stakeholders in the debate. The identified risks associated with AI in the media encompass disinformation, hybrid warfare, empathic media, deep fakes, targeted advertising, hate speech, and fake accounts. While the EU's AI Act and the Council of Europe's Recommendations adopt a risk-based approach and emphasize digital literacy, data protection, transparency, and collaboration with media regulators and consumer associations, there remain many open questions about the regulation of specific AI technologies used in the media and the protection of consumer rights within the media landscape.
The author concludes that, firstly, there is a need to address the regulatory focus on platform algorithms, which tends to neglect the impact of AI tools within news media. Secondly, finding a balance between transparency and user experience becomes crucial at the intersection of fundamental rights. Lastly, there is a pressing need for self-regulation and the establishment of ethical guidelines in newsrooms to tackle the increasing automation of journalism and the ethical implications associated with AI technology.

Platform regulation and "overblocking" -The NetzDG discourse in Germany
In their text, Jens Pohlmann, Adrien Barbaresi, and Peter Leinen examine the discussion about the Network Enforcement Act (NetzDG), the German anti-hate speech law. They analyze hundreds of articles from December 2016 to August 2019 on tech websites (such as Netzpolitik and Heise) and tech sections of German newspapers (such as Frankfurter Allgemeine Zeitung and Süddeutsche Zeitung) about NetzDG.
Methodologically and empirically, this article is innovative as it uses transformer-based language models to detect recurrent arguments in the texts, enabling subsequent close reading. As automated processes, language models cannot 'understand' any specific idea, but they can certainly assist researchers in the quantitative analysis of arguments. Following the trend towards open science, the authors make the whole dataset of German tech websites available free of charge.
The authors have found out that the "overblocking hypothesis" is predominant in the discourse both on tech websites and newspapers. This hypothesis maintains that the short time limits to decide about unlawful content lead platforms to delete or block more content than necessary. Although the article concentrates exclusively on the discourse around NetzDG, the same argument has been used in nearly all Western countries experimenting with platform regulation on speech, including the discussion around the Digital Services Act (DSA) -hence the broader relevance of this approach.

The "neo-intermediation" of large on-line platforms: Perspectives of analysis of the "state of health" of the digital information ecosystem
Europe, it is assumed by Isabella de Vivo, has a fundamental problem with the proposed way to regulate harmful content on digital platforms. She develops the argument that the currently applied regulatory mechanisms are not appropriate for handling this complex issue, currently protected by the (US-American) safe harbor principle. In her contribution, the author develops the idea of extending "re-intermediation" further into "neo-intermediation" as a third way between state regulation and self-regulation. She includes users in the accountability design and suggests implementing specialized bodies such as independent intermediaries between regulators and platforms to assist producers at their own request.
While addressing a critical issue and using recommender systems as well as the DSA as cases to exemplify her idea, such 'triangulation' might create a new layer of complexity with uncertain consequences. Nonetheless, this text contributes to the ongoing debate on how to 'cage' unwanted and harmful content on digital platforms, or indeed, how to cage digital platforms by regulation altogether.

Where to next with Australia's news media and digital platforms Mandatory Bargaining Code?
The special issue concludes with an analysis of platform regulation from outside Europe that influences much of the current EU discussions. Australia made platform policy headline news with its announcement to implement a Mandatory Bargaining Code aimed at addressing the fundamental power imbalance between digital global platforms and the national news media businesses. This move was expected to benefit major news media companies by providing a consistent revenue stream diverted from the two dominant platform companies, Alphabet (Google) and Meta (Facebook). Tim Dwyer, Terry Flew, and Derek Wilding note in their analysis that the ultimate beneficiaries and the amount of money generated by the Code remain uncertain or unknown. The figure currently in question is 200 million AUD annually (120 million euros).
Upon closer examination, the Code reveals significant flaws, including weak bargaining power for small and regional newspapers that lack secured deals. Eligible news media are those whose main purpose is to produce democracy-enhancing, public-interest journalism. They have to bargain their deal with the Australian entities of Alphabet and Meta. Above all, the Code exists only as a threat that is yet to be activated, Dwyer and colleagues report, as most stakeholders have managed to settle agreements so far.
The Code has sparked interest worldwide as a potential way to rebalance at least to some extent the power structures that have put the established news companies at a significant disadvantage vis-à-vis digital platforms. It turns out, however, that these global giants have no real incentive to share their revenues, even while there is some willingness to respond to concerns over the spread of disinformation. The authors suggest several ways to address the Code's obvious shortcomings, such as giving the oversight of the bargaining processes to the regulator rather than the minister, being transparent about the fund's impact on the quality of public interest journalism, and establishing robust enforcement actions.
Given the Code's complex architecture from bargaining to arbitration, it might be worth considering a Google representative's proposal to replace the Code by a government-run industry fund fed by a levy payable by digital platforms. Such a fund would mimic media subsidy schemes that have been operating successfully in numerous European countries for decades.

Conclusion
Our special issue contributes to the debate on European media and platform policy by providing a more nuanced approach to the problems associated with the platforms. This does not mean that platforms are not to blame for problems in the contemporary media system, but that a proper understanding of platforms' more specific roles in these problems will highlight the kinds of interventions that are more likely to succeed.
Regulating platforms for public-interest oriented communication involves proper assessment of the overlapping roles played by all stakeholders, and not only the digital platforms, in the contemporary media systems (Humprecht et al., 2020;Pickard, 2020). Platforms should act in a more responsible way with regard to content, as shown by the articles of this special issue, but their dominance over other social and communicative actors, such as the media, must also be addressed (Helberger, 2020). If this does not happen, there is the risk of excessively focusing on the responsibility of platforms to govern content and, as such, entrenching the market power of a few entities across the whole European Union (Bayer, 2021).
In our view, this requires abandoning techno-exceptionalism in order to acknowledge that, despite all the hype, digital platforms are companies operating communication infrastructures as well as media content and playing within the rules (even if at their margins) of sovereign political systems (Chenou, 2014;Griffin, 2023). These political entities, including the European Union and its member states, have the conditions to design new rules in the public interest, and there is a rich history of successful attempts to do so -for both infrastructures and media content (Nielsen et al., 2019;Rahman, 2018). From this point of view, the main problem is how to build the political strength to reshape and regulate digital communication in the public interest. In our opinion, after understanding the development of the field in the past three years, this should be the next step of research on media and platform policy.
As guest editors of this special issue, we express our gratitude to all authors who delivered their manuscripts according to peer-review recommendations, to our reviewers, to our fellow scholars participating in the Jean Monnet Network EuromediApp and, first and foremost, to the highly competent team of Communications who guided us smoothly through the editorial process.