On 17 February 2021, Facebook stopped its Australian users from sharing local or international news. In addition, Australian news organizations that operated Facebook Pages were not only unable to share content on their pages but also had their Pages wiped of historical content. Facebook had been arguing with the Australian Government over a controversial new law that would force platforms to pay for news content and had regularly threatened to withdraw services from the country as the policy was being developed. However, it was still something of a surprise when the ban happened. Overnight, Facebook stopped serving news to an entire country and no-one could stop them. They only reinstated news after the Australian Government made last-minute concessions (Meade, Taylor and Hurst 2021).

Facebook has a history of making unilateral decisions around how news is distributed on their platform. In January 2018, the company decided to prioritize content from family and friends over content from Facebook Pages across its entire platform (Wong 2018). However, news companies across the world used Facebook Pages to distribute stories to their audiences. The change had a significant impact on outlets that focused heavily on social distribution, forcing them to diversify their distribution strategies (Meese and Hurcombe 2020). The change may also have had political impacts. Some have reported that Facebook tweaked its algorithm to favour right-wing content after finding at the testing stage that a significant amount of right-wing news content would be removed under the new rules. Mother Jones reported that some left-wing outlets (including their own) were targeted to ensure that more right-wing content circulated after the algorithm change (Bauerlein and Jeffery 2020).

Google makes decisions in a similarly autonomous manner. The company distributes algorithmically curated news through Google News and delivers news content occasionally through Google Search. In some countries, a “Top Stories” carousel appears at the top of relevant searches. The company has been careful to build partnerships with publishers and at one stage both sectors were working on the Accelerated Mobile Pages (AMP) project, which would see news content delivered quickly and seamlessly to people’s smartphones. However, the news sector soon accused Google of focusing on their own interests and prioritizing publishers who used AMP over competitors when selecting stories to appear at the stop of search results (Scott 2018). Google admitted to preferential treatment in late-2020 and announced they would no longer focus on the use of AMP as the sole decision point when working out which stories would appear in the coveted carousel positions (Jeffries 2020).

These three examples give us some insight into the problems that start to appear when platforms engage in news distribution. They use inscrutable algorithms to make significant decisions around the visibility of news on their service, which can affect people who want to access news and news outlets who want audiences to reach their websites. As a result, these technology companies are not merely intermediaries or conduits but have become critical gatekeepers (Blanchette 2021; Jeffries 2020; Napoli 2015; Russell 2019; Wallace 2018). When they make decisions, there is often little that the media industry can do about it. Thankfully, platform regulation has started to become a critical policy issue.

Media activists and certain scholars have been concerned about the rising power of digital platforms since the early 2010s (Daly 2016; Fuchs 2013; Lovink 2011). However, the international policy community only began paying serious attention to the problems associated with the dominance of Google, Amazon, Facebook and Apple (otherwise known as GAFA) from the mid-2010s onwards (Galloway 2017). Policymakers and legislators started to focus on market concentration (Kahn 2016; Moore and Tambini 2018) and related issues like tax avoidance (British Broadcasting Corporation 2020), privacy (Solove 2004; Cohen 2013), misinformation (or fake news) (Farkas and Schou 2019; Tandoc Jr., Lim and Ling 2018) and, critically for this volume, the algorithmic distribution of news.

Wider regulatory interest in the governance of the algorithmic distribution of news arose as platforms and news outlets developed an increasingly strained relationship across the 2010s (Bell and Owen 2017). At the beginning of the decade, many news outlets these new intermediaries as a potential solution to their economic problems. While print advertising revenue had dried up, news media companies figured that they could build larger audiences through social media and direct these new people to their online websites (Bossio 2017; Usher 2014). They could then generate income by charging advertisers more money to advertise on their increasingly popular websites. While some outlets were more cynical of these new intermediaries, Google and Facebook were keen to partner with news organizations and offered funded partnerships to various outlets. Initially, the relationship was working out. Platforms got a steady stream of professionally produced content and news outlets saw a significant increase in traffic to their websites (Tandoc Jr 2014; Nielsen and Ganter 2018; Zamith 2018).

However, a breakup was on the horizon. Platforms started to capture an increasing amount of digital advertising revenue, which meant that there were not many dollars left for news companies (Australian Competition and Consumer Commission 2019). Opaque and unpredictable algorithms also caused major problems for news organizations (Christin 2020; Meese and Hurcombe 2020). Editors and journalists struggled to gain visibility on these platforms. They made significant investments at the behest of platforms, increasing their production of video content after Facebook signalled that they would prioritize video (Tandoc Jr and Maitra 2018). However, platforms would then change their mind and leave media outlets in the lurch. The relationship became increasingly tense and news outlets in various countries (most notably Australia), started demanding regulatory intervention (Meese 2020). The media sector suggested that the public interest outcomes associated with the distribution of news were being harmed by this increasingly unworkable relationship. They also argued that platforms were essentially stealing their content and called for news organizations to be paid for the snippets of content that appeared on search results or on Facebook News Feeds (Flew and Wilding, 2020; Meese, 2020).

Further impetus for reform came from policymakers, legislators and scholars who were concerned about what the convergence of news, platforms and algorithms meant for democracy. Initial worries were about the prospect of “filter bubbles” and “echo chambers” appearing. These terms were coined by Eli Pariser (2011) and Cass Sunstein (2018), respectively, in response to the growing personalization of online content. Pariser worried that algorithms might only deliver a limited selection of news to people, reinforcing existing beliefs, and Sunstein was concerned about the homogenization of political discourse across social networks. Eventually, Axel Bruns (2019, 8) showed that people had much more diverse media diets than these concepts suggested, which themselves were based on “hypothetical thought experiments or personal anecdotes”. Nevertheless, the potential of platforms to unilaterally decide how and when people would access news and other content was still of concern. As a result, policymakers started to focus on algorithmic transparency and the operation of recommender systems from the mid-2010s onwards.

The emergence of these two related policy trajectories has seen countries across the world propose or implement major reforms, which specifically address the algorithmic distribution of news. The collection grapples with this moment of reform, focusing on countries outside of the United States. This geographic distinction is important for two reasons. Firstly, much of the regulatory concern and activity is occurring in countries who have to respond to the rapid growth of US-based platforms (Meese 2020). The United States, on the other hand, has generally taken a relaxed approach to platform regulation until the late-2010s, whereas other jurisdictions have been actively considering reform for a longer period of time. Secondly, algorithmic distribution is not just an issue for democratic countries. For example, Chinese platforms also distribute news in this way, but their political system ends up producing somewhat different regulatory interventions. This international approach allows us to provide a comprehensive discussion around the different ways that countries have responded to algorithmic distribution as a social phenomenon and a policy problem.

The chapters in this volume cover the policy responses of a diverse set of countries to the algorithmic distribution of news—a transformation in news business models led, in a large part, by American multinational platforms. We examine policy responses to these changes from around the world—in Europe (Chapter 7 by David Lindsay, Chapter 8 by Christian Herzog, Christopher Buschow, and Alessandro Immanuel Beil, Chapter 15 by Jannick Kirk Sørensen, and Chapter 14 by Judith Vermulen), including Switzerland (Chapter 12 by Colin Porlezza) and Germany (Chapter 11 by Kerstin Liesem; Chapter 8 by Christian Herzog, Christopher Buschow, and Alessandro Immanuel Beil); in the Asia Pacific region, including New Zealand (Chapter 10 by Merja Myllylahti), Australia (Chapter 7 by David Lindsay; Chapter 9 by Tai Neilson and Baskaran Balasingham; Chapter 6 by James Meese; and Chapter 4 by Catherine Young), and China (Chapter 2 by Jian Xu and Terry Flew; in North America, including Canada (Chapter 3 by Nicole Blanchett, Fenwick McKelvey and Colette Brin); and in Sub-Saharan Africa in Kenya (Chapter 5 by George Ogola and David Cheruiyot). We find commonalities across many of these regions—particularly the tendency in some countries to follow the regulatory lead of more powerful countries—as well as differences, including differences in the policy tools chosen to approach the algorithmic distribution of news.

The rest of our introduction proceeds as follows. We canvass these regulatory trends and outline some of the more popular conceptual responses. We go on to discuss historical institutionalism, the theoretical and methodological approach that informs the collection as a whole. After this summarize the major themes from this collection, before introducing each chapter and ending with a reflection on future research directions for journalism and media policy scholars.

Algorithmic Governance: Transparency, Diversity or Departure

As part of this wider regulatory pushback, regional groupings and individual countries are actively working regulate recommender systems and other forms of algorithmic distribution. One approach involves introducing specific regulations to provide transparency about how these systems work. The European Union and its member states are deciding whether or not to adopt a proposed Digital Services Act (DSA), a wide-ranging reform that aims to regulate large online intermediaries across a variety of domains, from content takedowns to online advertising (Helberger 2021). Alongside these areas, the proposed reform also aims to give individuals transparency around recommender systems through Article 29. The Act would require specified platforms to make public what “the main parameters of their recommender system are and the options for users to modify or influence those parameters” (Helberger 2021). However, as Natali Helberger (2021) and colleagues explain, transparency does not necessarily mean control. Platforms may not be required to provide these options and moreover, even if they do so, individuals may be forced to choose between two or three subpar options, none of which may be aligned with “broader public and societal values” (Helberger 2021).

This turn towards transparency is evident in other jurisdictions as well, most notably Australia. In this jurisdiction, the major outcome of this recent international reform moment has been an inquiry around digital platforms, which Chapters 6, 7 and 9 in this volume discuss in more detail. The first major proposal to emerge from the inquiry was the News Media Bargaining Code (NMBC), which forced platforms to form commercial agreements with news outlets around the use of their content (Flew, Gillett, Martin and Sunman 2021). However, somewhat hidden in the Act that featured this headline reform were a number of additional standards that designated platforms would be required to follow.Footnote 1 One such standard required platforms to give news outlets advance notification of algorithm changes if it affected referral traffic to news content. In contrast to the DSA, this form of transparency is situated within the broader corporatist arrangement of the NMBC and as a result, specifically focuses on how algorithms impact on one industry. This corporatist arrangement is also evident by the fact that platforms and publishers can agree to “contract out” of any or all minimum standards (i.e. agree that they will not apply). The UK and Canada have watched the Australian reform process closely and are considering introducing similar or related reforms (Meese, 2020).

Other interventions address algorithmic distribution from the perspective of media diversity. The Australian reform process includes a non-discrimination provision that requires designated platforms to not choose between news outlets based on the outcome of commercial negotiations or other external factors (Australian Competition and Consumer Commission 2019). Germany’s Interstate Media Treaty (discussed in this volume) has a similar reform that prohibits large platforms from unfairly discriminating between news outlets (Nelson and Jaursch 2020). These are important reforms that start to treat platforms as public infrastructures. Their content becomes subject to state regulation, which aims to ensure that people get a diverse selection of news delivered to them. Leaving aside the larger debates around the ongoing focus on content published across organizations (external diversity), at the expense of diversifying content published within organizations (internal diversity), these reforms do not actively attempt to impose diversity requirements on platforms (Karppinen 2013; Loecherbach, Moeller, Trilling and van Atteveldt 2020). The aim is to ensure that no outlet is unfairly excluded on these platforms, rather than genuinely understand what a diverse media selection on social media could look like. Chapters 14 and 15 weigh up the value of more interventionist approaches to the diversity problem.

Other approaches are either more idealistic or ideological. In the idealistic category, we could consider the growing interest building publicly funded services to replace their commercial equivalents. For example, instead of forcing Google to follow certain publicly oriented outcomes, could a country (or perhaps, the world) just publicly fund a search engine? Related outcomes associated with the public interest (Napoli, 2019), like securing media diversity, might be achieved with less friction through this approach. Existing work on the topic suggests that “public service models for search engines – and especially social networking – would likely function poorly if made available only on a national basis” (Andrejevic 2013, 131) but suggests that international co-operation could be a possibility. Scholars have even argued that “nonmarket provision” has become even more critical in the current climate because existing policymaking efforts are failing to genuinely grapple with the fact that platforms are fundamentally opposed to democratic outcomes (Cammaerts and Mansell 2020, 147). Chapter 13 engages with these questions and considers whether an open-sourced social media platform, supported by public service media could break through the current state of platform dominance.

Conversely, in the ideological category, we have a diversity of policy responses that are more politically inflected. Perhaps unsurprisingly, the two standout nations in this regard are the United States and China. The United States has been hesitant to regulate platforms in any domain, let alone in reference to the algorithmic distribution of news. This is partially due to the fact that the leading transnational platforms are based in this jurisdiction, and the nation has had a vested interest in ensuring their success (Popiel 2018). The country’s strong commitment to free speech through the First Amendment has also influenced its approach to online platform regulation. The Department of Justice (2020) has launched an antitrust suit against Google and Congress is considering an algorithmic accountability act, which would provide more oversight around automated decision-making systems (Algorithmic Accountability Act of 2019). However, there has been no sign that the Congress will regulate how algorithmic systems shape the circulation of news. China has taken the opposite approach and as one of our chapters will show, has taken an active role in this area. The relevant regulatory bodies organize regular meetings with the leading news aggregation service Jinri Toutiao (Today’s Headlines), while the Chinese state media has organized a competing product called Renminhao (Good People).

Through this brief survey of regulatory trends, we see that a number of countries are regulating platforms in an attempt to manage the algorithmic distribution of news. We have identified a number of different approaches, which as we will see throughout the collection, are often influenced by local political contexts and broader regional developments. This focus on nations and regions furthers another goal of this collection, which is to explore the geopolitics of media regulation. China is working to transform its nation into a platform and export this digital infrastructure across the world, while maintaining a high level of internal control (Choudary 2020). In contrast, the United States is touting its transnational platform companies as national success stories and tracking China’s growing tendency to exhibit its global strength through technological means (Weber 2018). Many of the countries in our collection are stuck in the middle of this tech-influenced geopolitical manoeuvring and are attempting to establish some independence from the United States, without subsequently becoming dependent on China. Our collection features a number of small (Switzerland, New Zealand) and middle powers (Canada, Australia) attempting to negotiate this difficult situation.

Major Concepts in News, Algorithms and the Public Interest

The scholarly debate around the algorithmic distribution of news has begun to mature as regulatory reforms have occurred. The vague fears and catchy phrases of Sunstein (2018) (“echo chambers”) and Pariser (“filter bubbles”) (2011) have given way to rigorous research projects and considered conceptual development. Journalism scholars have offered important insights. There has been significant research on the gradual introduction of algorithmic logics into newsrooms. Scholars have tracked the increasing use of analytics to make editorial decisions (Christin and Petre 2020; Tandoc Jr 2014; Petre 2015; Zamith 2018), the growing focus on distribution through digital platforms (Diakopolous 2019; Bossio 2017) and the importance of news aggregation systems and services (Coddington 2019). The field has also noted that these new intermediaries are gatekeepers who have as much capacity to select and feature news as traditional gatekeepers (like the editor of a newspaper) (Shoemaker and Vos 2009; Wallace 2018). Julian Wallace has gone so far as to propose a new theoretical model of gatekeeping to better address the complex distributional networks now in operation (Wallace 2018).

Alongside this work is a growing body of media policy scholarship, which this collection contributes to. Natali Helberger has established a comprehensive research agenda around news recommender systems that has made significant empirical and theoretical contributions. Alongside her colleagues, she found that in certain circumstances algorithmic recommendations can mimic the sort of news diversity found in recommendations made by human editors (Möller, Trilling, Helberger and van Es 2018). Her other work has offered an important critical perspective on algorithmic recommendations, explaining that these technical advances could have positive and negative implications, depending on how one understands democracy. She suggests recommenders that emphasize facilitating individuals’ management of their information diets, and recommenders that focus on presenting information readers “ought to read” to “alert, inform or even educate readers and push them out of their intellectual comfort zones” can both be compatible with democracy (Helberger 2019, 1010).

Philip Napoli (2019) has also made an important foundational contribution in this area. He (along with Robyn Caplan) argues that social media platforms need to be treated as publishers (essentially comparable to other media outlets) and actively regulated in the public interest (Napoli and Caplan 2017). It is clear from this our introduction and contributions throughout this volume that governments have followed this directive only in the most general sense. They have approached platforms and their algorithms as regulatable objects but have not established a complete equivalence between platforms and publishers. Many of these reform agendas also appear to align with the goals of the commercial media industry or interest groups across the cultural policy sector. This speaks to a wider historical problem across media policymaking, which sees powerful actors prioritized over civil society and the wider public (Flew, Gillett, Martin, and Sunman 2021; Freedman 2008).

Indeed, many of our featured jurisdictions appear to have engaged in relatively light-touch regulation when viewed against publicly oriented reforms proposed by scholars (Helberger 2019) or proposals of non-market alternatives (Andrejevic 2013; Cammaerts and Mansell 2020). Other researchers question whether the dominant regulatory approaches we have identified above are even taking the right conceptual approach; Mike Ananny and Kate Crawford (2018) ask if growing demands for algorithmic transparency are adequate to ensure algorithmic accountability. They argue that regulators and policymakers need to adopt a systemic approach of looking at the operational logic of algorithmic systems and identifying who is involved as an actor, who has the power to intervene and whether the system is legible enough to be deployed, amongst other things. These alternative approaches present useful insights for those countries in our collection that are still considering reform and underline the fact that all interested parties have not converged around one ideal model to regulate algorithms, platforms and the public sphere.

Historical Institutionalism: Path Dependency, Change and Critical Junctures

The above discussion of power, gatekeeping and the public interest leads us to historical institutionalism, the conceptual framework that sits at the heart of our collection. Historical institutionalism sits within a broader set of “new institutionalist” approaches (March and Olsen 1983) that aim to account for the role of institutions across society. As one of us has previously noted, communication scholarship has engaged with institutional theory in an ad-hoc manner and “tend[s] to speak of new institutionalism without explicitly distinguishing one institutionalism from another” (Bannerman and Haggart 2015, 2). If more specificity is provided, scholars tend to focus on sociological institutionalism, an approach that focuses on cultural practices within organizations (DiMaggio and Powell 1983). Existing journalism scholarship has used this framework to understand how changing newsroom practices affect the institution of journalism (Anderson, 2012; Benson, 1999; Benson, 2004; 2006). It has also been used to explore how Facebook’s algorithmic logics have intersected with journalism (Caplan and boyd 2018).

In contrast, historical institutionalism focuses on the history of each institution as playing a defining role in its subsequent development. The role of history is accounted for through the concept of path dependency, the idea that “past decisions tend to constrain future institutional change” (Bannerman and Haggart 2015, 5). The approach rejects the claim that the future is radically unknowable and instead argues that previous institutional decisions inform and shape future outcomes. Another core concept is the idea of the critical juncture. While institutions are generally viewed as path dependent and stable, at important points there is a moment of disruption where institutional arrangements are uncertain, and change is seriously considered (Capoccia 2015). The collection views the growing importance of platform regulation and the growth of algorithmic distribution as critical junctures. There is tension between acknowledging the path dependencies of history and examining change (which is also rooted in past fissures and tensions). Critical junctures can provide moments of possible agency, even as actors are never fully free from existing path dependencies. Instead, these moments either reinforce institutions (and associated path dependencies) or weaken them (Bannerman and Haggart 2015). Across our chapters, we can see how different countries and journalism more broadly have responded to these moments and assess the extent to which different actors can actually make radical departures.

Importantly for this collection, historical institutionalism also allows scholars to focus on the structural relationships between different institutions. This is because it accounts for power relations as part of its overall approach and places a greater focus on political contests (Bannerman and Haggart 2015). This makes it arguably suited to studies of policy reform, opposing interests and difficult moments of institutional change. The collection features contributors from a variety of disciplinary backgrounds; not all of our contributors specifically adopt historical institutionalism. However, each chapter is attuned to the wider concerns of this approach, recognizing the importance of historical developments and accounting for moments of continuity and change.

Change and Path Dependency: The Findings of This Volume

The chapters that follow reveal both significant change and significant path dependency in news production and related policies. Markets and business structures have undergone change and disruption, as Meese, Porlezza, and Young note in this volume. The emergence of players like Facebook and Google News has significantly changed and disrupted advertising markets and the business models of journalism organizations and news publishers. This emergence has also brought structural changes to news consumption practices, production processes, and to newsrooms’ infrastructures, interactions with readers (Young) and distribution practices (Porlezza; Young). More broadly, these changes are tied to changes across media systems and to a shift, as Hrynyshyn notes, from mass media to a more individualized media system.

A raft of significant legal and policy changes, from the privatization of telecommunications (Ogola and Cheruiyot) to the advent of artificial intelligence policies (Porlezza, Myllylahti), lie behind and respond to the rise of algorithmic news distributors. Chapters in this volume address the variety of policy realms that are being adapted in response to the algorithmic distribution of news, from copyright (Lindsay; Herzog, Buschow, and Beil; Liesem; Meese), competition policy (Lindsay; Neilson and Balasingham, Meese), hate speech (Liesem; Myllylahti), media diversity (Sørensen, Vermullen), infrastructure policy (Hrynyshyn), and media licensing and censorship practices (Xu and Terry Flew). Insofar as the policy changes at hand regulate the gatekeeping functions (Blanchette, McKelvey and Brin), production and distribution of news, they arguably regulate structures that are of “systemic importance” for democracies (Gersemann 2019 quoted in Herzog, Buschow, and Beil) and authoritarian countries alike. Chapters in this volume stretch from covering the recent histories of policies that are currently in place, to speculative analyses of policy changes of the potential future—those that might govern the localization of platform infrastructures (Hrynyshyn) and media diversity (Vermulen, Sørensen).

Algorithms themselves also undergo change, as when Facebook altered its news feed algorithm following the Cambridge Analytica scandal, as Merja Myllylahti and James Meese discuss in this volume. Such changes can have dramatic effects on some news companies (Myllylahti, Meese) and, potentially, on the relative promotion of political perspectives (Meese). Future potential recommender and news distribution technologies could bring cause further changes to the algorithmic distribution of news (Vermulen, Sørensen).

While change may be a constant fact of life, the authors in this volume identify several changes as critical junctures. Some critical junctures are brought by events external to policy; the Christchurch mosque terrorist attacks in March 2019 (Myllylahti), the COVID-19 pandemic (Meese), and the Capitol riots of January 2021 (Hrynyshyn) were—each in different ways—moments of important rethinking about the role and effects of social media platforms in distributing news along with hate, terrorist communications and misinformation. Others are internal to the course of law and policymaking itself—new approaches introduced as past approaches break down, are undermined, or become “no longer viable” (Herzog, Buschow, and Beil). These include the introduction of a new ancillary copyright, which then drove “copyright reform across the EU and elsewhere” (Herzog, Buschow, and Beil); the German Interstate Media Treaty and NetzDG, which departed from previous regulatory approaches (Leisem); and the Australian News Media Bargaining Code (Neilson and Balasingham), now influential as a possible model for other countries, to name a few.

Despite a number of approaches that bring significant or foundational change, there are also significant path dependencies not only in law and policymaking, but also in media industries, journalistic practices, and business models. Established and powerful actors—including platforms themselves—sometimes seek to maintain the status quo if existing frameworks serve their interests (Herzog, Buschow, and Beil). Existing institutional arrangements and business models can hold back policy change (Meese). Because existing laws and policies serve powerful political and business interests, these can prove difficult to shift, particularly if political priorities remain aligned with existing laws and regulations (Xu and Flew; Myllylahti). Business models (like the news media’s reliance on advertising) can also become entrenched, along with attendant regulatory systems. These, along with barriers to entry for new players (Herzog, Buschow, and Beil) and the dependencies of longstanding players on existing business ecosystems (Meese), can prove to withstand—for better or for worse—significant technological and economic change (Ogola and Cheruiyot; Meese). Journalistic practices also, in many cases, prove at least partially resilient to the changes brought by new technologies and distribution practices (Young; Ogola and Cheruiyot; Blanchette, McKelvey, and Brin).

Chapter Overview

In section one, “In the newsroom: algorithms, bots, business models, and privacy”, Jian Xu and Terry Flew give a detailed review of the Chinese Government’s response to the algorithmic distribution of news and the incorporation of algorithms into its existing policies of Internet censorship and regulation. Nicole Blanchette, Fenwick McKelvey and Colette Brin discuss the ways that algorithms and social media have changed news production and distribution across Canadian newsrooms and outline subsequent policy responses. Catherine Young discusses the use of a chatbot by the Australian Broadcasting Corporation, and George Ogola and David Cheruiyot review the limited use of algorithms and metrics by Kenyan news media. These Canadian, Australian and Kenyan case studies emphasize the national policy contexts in which the algorithmic distribution of news is situated, underlining the need for government policy to be revised in light of these practices—particularly in relation to the collection of personal information by newsrooms in the course of conducting chatbot or analytic operations. James Meese, focusing on Australia, sheds light on the difficulties news organizations have had in responding to the algorithmic distribution of news with new business models, and the potential inadequacies of government responses under the New Media Bargaining Code that has recently been established.

In section two, we review the two leading policy approaches that have emerged to the algorithmic distribution of news: copyright reform and competition law reform. David Lyndsay compares the two approaches, arguing that the competition law approach is better suited to the policy problem at hand, focused on market power. Herzog, Buschow and Beil argue, in a review of the major statements of stakeholders, that a private property vision of copyright reigns in the establishment of ancillary copyright—or a right of publishers to receive copyright payments for the use of news snippets by news aggregators like Google News. Tai Neilson and Baskaran Balasingham give a helpful review of the centrality of competition law in Australia’s recently implemented reforms, and ask whether competition law could be extended to other areas of media policy, most notably media pluralism.

In section three, chapters focus on some of the challenges entailed in mounting regulations to respond to the algorithmic distribution of news. Merja Myllylahti reviews the challenges faced in New Zealand, which has not extended its active policy response following the Christchurch mosque terrorist attacks around hate speech and radicalization to other areas of platform regulation; Kerstin Liesem reviews the German approach to platform regulation, and Colin Porlezza discusses the challenges that Switzerland faces in studying and ultimately responding to platform regulation in light of the significant moves made by powerful neighbours like Germany. Common themes seen here are the need for less powerful countries to follow the regulatory trends of their more powerful trading partners, particularly in light of an overall reluctance to regulate social media in light of historically liberal media policies.

In the final section, we discuss some possible future regulatory areas. Derek Hrynyshyn proposes a radical break with social media infrastructures and current regulatory policies intended to address the fundamental public policy problems posed when asking—or allowing—profit-seeking platforms based on complex algorithms to govern speech. In the next two chapters, Judith Vermulen and Jannick Kirk Sørensen review the promises, possibilities and perils of designing or regulating news recommenders to encompass the policy objectives of exposing users to a diversity of views and content.

Future Directions: The Geopolitics of Communication Policy

The chapters in this volume reveal the growing importance of nation states and regional groupings in policy reform. Communications policy researchers have challenged the excessively utopian narratives of globalization and have acknowledged that the state continues to play a critical role in policy reform and implementation (Flew, Iosifidis and Steemers 2016; Flew and Waisbord 2015). An adjacent strand of research has pointed to the emergence of a potential Splinternet (Malcomson 2016; Hoffman, Lazanski and Taylor 2020). With Europe, China and the United States regulating the Internet in vastly different ways, some suggest that these underlying legal frameworks could see the online world separate into “the authoritarian Internet, the bourgeois European Internet [and] the commercial American Internet” (Nabben, Poblet and Gardner-Stephen 2020, 747; O’Hara and Hall 2020). Throughout our collection, we see smaller and middle powers respond in a variety of ways. For example, Switzerland are policy followers and wait to see what larger states in their region do, whereas Australia has worked to establish themselves as international policy leaders. Powerful nation states (China) and regional groupings (the European Union) are very much able to act independently and take a more interventionist approach.

In response to this growing regional divergence amongst stronger actors, we are seeing a growth in international connections and collaboration. Examples include the launch of the International Grand Committee, an ad-hoc group of parliamentary representatives from around the world who meet to discuss technology policy, and the Christchurch Call, a New Zealand led initiative to prohibit the circulation of terrorist content online. We also see more informal and gradual efforts at policy translation, with the Canadian government closely watching the progress of an Australian reform that forced platforms to pay for news, a policy idea that was first tested in Europe (Meese 2020). These trends have given some people hope that there may be potential for genuine multilateral responses to algorithmic power.

However, multilateralism is difficult to achieve in practice, even amongst smaller nations who are less inclined (and less able) to establish and prosecute their own independent regulatory agenda. In this volume, contributors discuss how algorithmic distribution intersects with media policy and adjacent areas like competition law and privacy. We see how each nation’s legal, social and cultural contexts produce vastly different regulatory approaches. Some countries do not provide public funding to the media (Pickard 2019) whereas others offer generous press subsidies (Ots 2009). Governments may view media pluralism in completely different ways and have contrasting feelings about the value of national content or whether regulators should be involved in mandating algorithmic diversity. This does not mean that effective multilateralism and cross-border collaboration are impossible. Instead, it is a genuine challenge, one that will need to be carefully addressed as future reform agendas are canvassed.

Our collection suggests several additional areas for future research. Most critically, this collection treats distribution seriously (Braun, 2019; Lobato, 2012) and places it at the centre of conversations around journalism and media policy. Journalism scholarship tends to focus on the production and output of news and so “much of the literature on media distribution comes from scholars focused on other areas of communication” (Braun 2019, 3). The growing importance of algorithmic distribution means that this often-unloved intermediate space, sitting between production and consumption, must be given greater attention. While exceptional work on recommender systems has already been conducted (Coddington 2019; Helberger 2019), more research is needed, particularly in light of these recent reforms. One research trajectory, worthy of further exploration is the question of whether people want more control and choice when using news recommender systems (Harambam, Bountouridis, Makhortykh and Van Hoboken 2019). Would they be interested in tweaking their algorithms, and would their changes even affect the wider sociotechnical systems of which they are a part (Ananny and Crawford 2018)?

The collection sets the stage for more targeted comparative research across a number of topics. We have offered a necessarily limited survey of recent developments and are writing this introduction at a critical juncture. As this volume came together, governments were introducing or actively considering regulations around the algorithmic distribution of the news. In a few years, long-term outcomes can be fully assessed and more substantive comparisons between countries can be drawn. Scholarship in this area can also provide more granular accounts of these reforms and attend to differences between specific countries. Similar comparative work will need to be conducted around other relevant media policy issues. Media diversity has come to the fore in this volume and as many before us have noted, the concept is still ill-defined (Karppinen 2013; Loecherbach, Moeller, Trilling and van Atteveldt 2020). It is unclear how specific countries will operationalize the term in an algorithmic context and divergence between nations is likely here. Finally, the survival of many journalism organizations is an ongoing issue. Countries and regions will take different approaches to public funding, commercialization and press freedom. Comparative work can help us understand the path dependencies that inform policy choices in different jurisdictions and the effectiveness of specific interventions.

In closing, it would be remiss of us not to acknowledge that this edited collection was written during a global pandemic, with the contract signed just weeks before the world was thrown into chaos. Our international group of collaborators had to face significant personal and professional challenges as this volume came together and we thank them for their commitment during this difficult time. Misinformation also circulated across social media platforms as people struggled to come to terms with the reality of the situation. Governments and international bodies were forced to react in real-time. They introduced education campaigns and encouraged platforms to surface authorized information about the virus. These developments point to the ongoing problems that surround algorithmic distribution more generally and suggest that it will be an active policy issue for some time to come.

We hope you enjoy the volume.