The platform beat: Algorithmic watchdogs in the disinformation age

Abstract

As digital platforms have come to play a central role in the news and information ecosystem, a new realm of watchdog journalism has emerged – the platform beat. Journalists on the platform beat report on the operation, use and misuse of social media platforms and search engines. The platform beat can serve as an important mechanism for increasing the accountability of digital platforms, in ways that can affect public trust in the platforms, but that can also, hopefully, lead to the development of stronger, more reliable, and ultimately more trustworthy, platforms. However, there are a number of tensions, vulnerabilities and potential conflicts of interest that characterize the platform beat. This article explores these complex dynamics of the platform beat in an effort assess the capacity of those on the platform beat to enhance the accountability and trustworthiness of digital platforms.

Keywords

Algorithms news organizations platforms search engines social media watchdog journalism

Introduction

The precise institutional identity of the large digital platforms that operate within contemporary news and information ecosystems remains ambiguous. While the application of an institutional analytical framework to digital platforms has become increasingly common (Caplan and boyd, 2018), the question of the appropriate institutional identity to apply continues to be contentious. The platforms themselves have waged efforts to craft techno-centric, utilitarian identities for themselves that work against any kind of more socially or politically grounded institutional conceptualization, in an effort to sidestep responsibilities that would accompany particular institutional identities (e.g. the social responsibilities attached to media institutions; Napoli and Caplan, 2017).

In the past few years, however, we have seen an important shift, as external and internal pressures have forced these platforms to think of themselves less as ‘tools’ for individual users and more as institutions with formalized governance structures and explicit social responsibilities (Napoli, 2019). Recent events, such as the coronavirus pandemic, seem to have compelled the platforms to move even further in this direction (e.g. Paul, 2020).

The largest digital platforms, such as Facebook and Google, have frequently been categorized as nation-states in their own right (e.g. Grygiel, 2019; Ingram, 2010). Facebook CEO Mark Zuckerberg has said that ‘In a lot of ways, Facebook is more like a government than a traditional company’ (Kirkpatrick, 2018). This question of digital platforms’ institutional identity is intertwined with the focus of this article – how the behaviours and uses of these platforms fit within the watchdog function of the institution of journalism, in which news organizations serve as a check on the power of executive, legislative and judiciary bodies, as well as on other powerful societal institutions. If these platforms are anything approaching a quasi-governmental institution, then it is certainly important to evaluate how the dynamics of the relationship between these platforms and news organizations impact journalistic performance in what is termed here the platform beat.

The platform beat is an increasingly vital component of contemporary journalism. It involves the press holding digital platforms accountable for their actions, inactions and design choices, as well as for the actions of the users who operate within – and often exploit key characteristics of – these platforms. As Diakopoulos (2013) has noted, algorithmic systems, such as those that power search engines and social media platforms, represent a new form of power that can be exerted over citizens in a variety of contexts. And, as he also notes, ‘What we generally lack as a public is clarity about how algorithms exercise their power over us’ (Diakopoulos, 2013: 2, emphasis added). Napoli (2014) has gone so far as to classify algorithmic systems¹ as institutions in their own right, akin to other cultural, political and economic institutions. If we accept this position, then it would certainly make sense that these algorithmic systems, and the digital platforms in which they are embedded, become a focus of journalistic attention in the same way that journalism has served as a watchdog for other institutions.

However, the relationship between news organizations and digital platforms possesses some fundamental differences from the relationship between news organizations and other institutions that they cover. News organizations and digital platforms intertwine in ways that generate challenges for journalism’s watchdog function. Further challenges arise from distinctive characteristics that are inherent in contemporary platforms and contemporary news organizations.

The goal of this article is to explore how the complex inter-institutional dynamics between platforms and news organizations, along with fundamental characteristics of the platforms and contemporary news organizations themselves, might affect the capacity of news organizations to effectively cover the platform beat. Given the expanding role of large digital platforms in political life, this beat is increasingly important to the dynamics of platform governance and, thus, to a well-functioning democracy.

Delineating the platform beat

Journalists on the platform beat report on the operation, use and misuse of social media platforms and search engines. Specialized ‘technology reporters’ are increasingly common within news organizations. In addition, in many countries, specialized news operations have begun to emerge that focus explicitly on reporting on algorithmic systems and digital media platforms. For instance, in Canada, the Globe and Mail operates a tool that monitors and analyzes political advertising on Facebook (Owen, 2019). In Germany, Spiegel Online has established a reporting team dedicated to algorithmic accountability that has investigated topics such as personalization of Google search results and targeted advertising on Facebook (Elmer, 2020). In the United States, Craigslist founder Craig Newmark recently gave US$6 million to Consumer Reports to support reporting on ‘digital products and platforms’ (Tracy, 2019). This donation followed a US$20 million donation in 2018 to support The Markup, an investigative news operation that focuses on the tech sector and algorithmic systems (Tracy, 2019).

This notion of the platform beat builds upon previous areas of reporting emphasis such as ‘the algorithms beat’ (Diakopoulos, 2013), which focuses more broadly on algorithmic decision-making systems across all sectors of life (see also LaForme, 2018). The algorithms beat includes, for example, the role of algorithmic systems in criminal justice decisions, healthcare and finance. In contrast, the platform beat is more narrowly focused on the use and operation of the digital platforms that play an increasingly central role in the dissemination or curation of news and information. This focus reflects the distinctive role that these algorithmically driven platforms perform in the news and information ecosystem, and how their actions and uses can affect an informed citizenry in ways that other algorithmic systems, in areas such as criminal justice, and finance, do not.

This platform beat has two inter-related components. The first involves monitoring and scrutinizing the operation of these platforms, in terms of how they handle user data, how their advertising systems operate, and, of course, how their curation algorithms and content moderation systems perform. The second involves monitoring the content that is distributed on these platforms, looking for examples of widespread inauthentic behavior² and systematic disseminations of hate speech, calls to violence, disinformation and junk news.

In this regard, these news organizations are essentially policing the platforms’ performance as news and information distributors. And, given the increasingly central role of digital platforms in the news ecosystem, the platform beat involves journalistic oversight of institutions that are simultaneously (and, somewhat ironically) one of the most significant distributors of journalism. This dynamic essentially involves a tension between the computational-corporate model of trust (predicated on personalized data flows, algorithmic computation and proprietary business models) and the institutional–professional model of trust (predicated on human-made rules of power governed by publicly accountable institutions and professionals) that are the focus of this special issue (van Dijck, in press). Specifically, news organizations apply the traditional institutional–professional model of trust to their oversight and evaluation of these digital platforms – specifically, evaluating how well the platforms are performing in the application of their evolving technical–industrial model of trust to the content that they host and distribute. The irony here (and source of tension) is that the platforms’ application of this technical–industrial model of trust to their content curation and moderation processes encompasses these and other news organizations, which rely upon the platforms for distribution and audience engagement.

As a New York Times editorial noted, ‘Social media misinformation is becoming a newsroom beat in and of itself, as journalists find themselves acting as unpaid content moderators for these platforms’ (Editorial Board, 2018). As this same editorial noted, news organization reporting often serves as the means by which the platforms themselves become aware of flaws or biases in their operation, or of ways in which they are being exploited by purveyors of disinformation or hate speech (Editorial Board, 2018). In this way, these news organizations often perform a valuable service for these platforms, assisting them in identifying behaviours that violate their terms of service, or pointing to practices that might necessitate changes in platform governance. Julia Angwin (2020), Editor-in-Chief of the platform watchdog journalism organization The Markup, has gone so far as to ask ‘Are journalists and researchers the real content moderators?’

Journalistic accounts have been central to a growing number of revelations about the use and abuse of digital platforms. These include the Facebook activities of the Myanmar military against Rohingya Muslims (Mozur, 2018); the explosion of disinformation on WhatsApp in Brazil (Tardáguila et al., 2018); coronavirus conspiracies and misinformation (Lytvynenko, 2020); and, most recently, the flood of disinformation in the aftermath of the 2020 US presidential election (Wakabayashi, 2020). As the New York Times emphasized, ‘The public knows about each of these incitements because of reporting by news organizations’ (Editorial Board, 2018).

In a study of algorithmic accountability reporting (of which the platform beat is a key component), Diakopoulos (2019) identified four ‘different driving forces for algorithmic accountability stories’ (p. 4). These are the following: (1) discrimination and unfairness, (2) errors or mistakes in predictions or classifications, (3) legal or social norm violations and (4) misuse of algorithms by people either intentionally or inadvertently. In addressing these and related topics, the platform beat can serve as an important mechanism for increasing the accountability and trust of digital platforms, and influencing their governance decisions. However, as the next section will illustrate, there are a number of unique inter-institutional dynamics, and core characteristics of platforms and news organizations, that complicate the relationship between news organizations and the platforms that they cover in ways that can potentially undermine the effectiveness with which news organizations cover the platform beat.

The complexities of the platform beat

In comparison to the relationship between journalism and other institutions (e.g. government, corporations), the relationship between journalism and digital platforms has some important and distinctive layers of complexity. This section explores this complexity, with an eye towards identifying those factors that can impede news organizations’ ability to effectively cover the platform beat.

Economic interdependence

Perhaps one of the most prominent and distinctive aspects of the platform beat is the extent to which it is built upon an evolving foundation of interdependence between news organizations and digital platforms. Interdependence is nothing new in the relationship between journalism and the institutions that it covers. For instance, news organizations have long been dependent on government for various forms of ‘information subsidies’ (Gandy, 1982) to help them in their work. Even in a largely commercial media system such as the United States, news organizations have even received more traditional governmental subsidies in the form of postage discounts for their printed publications and free access to valuable broadcast spectrum (Cowan and Westphal, 2010). And, of course, when we are talking about the public service media systems that are so prominent across Europe, government subsidies are far more extensive than mere discounts on distribution costs. On the other side of the equation, government bodies and individual politicians or political candidates have traditionally relied, to some extent, on news organizations to disseminate their messages and generate attention to their campaigns through news coverage.

When we look at the nature of the interdependencies between news organizations and digital platforms, certainly versions of the subsidies described earlier are still present. Social media platforms represent a massive information subsidy to news organizations, as news organizations have become increasingly reliant on monitoring social media feeds as an inexpensive means of identifying breaking news and story ideas (Coddington, 2019).

In addition, as the platforms have frequently argued, they provide a valuable form of distribution for news organizations, providing them with access to aggregations of audience attention far beyond what these organizations are capable of generating on their own. Conversely, news organizations have frequently argued that they provide valuable content that, to some extent, fuels audience interest in – and usage of – the platforms (News Media Alliance, 2019).

Reflecting this perspective, European policymakers have imposed regulatory models that require platforms to compensate news organizations for the content that they disseminate. Germany and Spain, for instance, adopted ‘link taxes’ in 2013 and 2014, respectively, which required aggregators such as Google News to compensate news outlets when providing links to their content. Analyses of both these European cases showed that losing Google News distribution caused substantial harm to news outlets, which explains why, in Germany, many outlets were willing to forego compensation in order to maintain carriage (Calzada and Gil, 2020). Such patterns suggest that news organizations are more dependent on platforms for distribution than the platforms are dependent on news organizations for content.

Efforts to address this asymmetric dependency have become more pronounced of late. In November 2020, French publishers became the first publishers in the world to sign government-mandated content licencing agreements with Google; and, more recently, members of the European parliament have expressed a willingness to amend the Digital Services Act and Digital Markets Act in ways that could reflect legislation passed in Australia that requires digital platforms to negotiate compensation with news organizations (see Napoliand Royal, in press).

Despite this asymmetric dependency, it is important to recognize that the platforms, for the most part, have not elected to walk away from what has certainly been a fraught and contentious relationship with news organizations. Indeed, platforms such as Facebook and Google have instead entered into contractual relationships with news organizations that involve financial compensation (Napoli and Royal, in press).

Google recently announced a US$1 billion partnership with European news publishers in connection with its new Google News Showcase product (Hutchinson, 2020). In connection with its recently launched news tab, Facebook has begun paying select news organizations for access to their journalism (Kafka, 2019a), a substantial shift from the company’s long-standing position of refusing to pay publishers for content accessible on its platform. Similarly, in 2019, Apple launched a subscription news service that involves revenue sharing with news organizations (Kafka, 2019b).

These initiatives build upon longer-standing financial relationships that some of the platforms have maintained with journalistic fact-checking organizations. Facebook, for instance, contracts with 50 independent fact-checking organizations from many different countries (Wilson, 2020). In 2017, Google began integrating the work of fact-checkers into its search returns (Kosslyn and Yu, 2017). While the financial terms of these arrangements are seldom made public, in those instances in which financial information has been made available, the payments from the platforms have been shown to account for nearly half of the total revenue for these fact-checking organizations (Usher, 2020).

Journalism has also emerged as a point of focus for the philanthropic efforts of some of the major digital platforms. According to recent research, Google and Facebook’s publicly disclosed contributions to journalism total over $700 million since 2016 (Usher, 2020). Google’s Digital News Innovation Fund has donated over 140 million euros to news organizations across Europe since its launch in 2015 (Blecher, 2019). Platforms such as Twitter and Facebook have made substantial philanthropic donations specifically directed at supporting coronavirus-related reporting (Lunden, 2020). Furthermore, in 2020, Facebook committed US$25 million in emergency grant funding to local news organizations in the United States adversely affected by the coronavirus pandemic (Brown, 2020). These are just a sampling of the extensive philanthropic connections between digital platforms and journalism.

In sum, these digital platforms have emerged as some of the most significant funders of contemporary journalism – ‘platform patrons’ according to Ingram (2018). Benton (2020) has offered a number of critiques of this development. As he contendss, the platforms generally want to play a very active role in determining which news organizations receive funding; they tend to want to fund new initiatives (to avoid the appearance of returning core revenues that they have syphoned off); and they are primarily concerned with the public relations benefits that accrue from such initiatives (rather than thinking about how best to rehabilitate the news ecosystem; Benton, 2020).

Also, once a news organization receives compensation from a platform for its content, or charitable support, this raises the question of whether this support might affect how rigorously and objectively that organization covers the platform (Usher, 2020). Furthermore, simply the prospect of such compensation could colour the coverage of news organizations that aspire to receive support from the platforms. These are issues that echo long-standing concerns about the relationship between news organizations and advertisers, in terms of whether the potential for, or receipt of, advertising dollars might colour how news organizations cover the behaviours of significant advertisers, or how they might alter their content in ways that prioritize advertiser interests over journalistic values (Baker, 1994).

However, the key difference in the scenario being discussed here involves the centrality of distribution to this relationship. Unlike advertisers, digital platforms’ governance decisions can dramatically and instantaneously affect the distribution of journalism (see, for example, Caplan and boyd, 2018) – a dynamic that puts news organizations in a very vulnerable position in terms of feeling empowered to perform aggressively on the platform beat.

Competition

Another key complexity in the relationship between digital platforms and news organizations is the extent to which they are competitors. News organizations and digital platforms compete directly for audience attention and advertising dollars. However, this competition has become exceptionally one-sided. It has been well-documented how large digital platforms such as Facebook, Google and Amazon have come to dominate the digital advertising market (Fischer, 2020). In so doing, these platforms have syphoned off some of the key categories of advertising that were once the lifeblood of news organizations. Platforms such as LinkedIn, for instance, have replaced print or digital newspapers as venues for employment advertising. Platforms such as Amazon and eBay have attracted much of the classified advertising dollars that once would have flowed primarily to newspapers.

This competitive dynamic between platforms and news organizations raises questions about how objectively news organizations might perform on the platform beat. Platform executives frequently bemoan what they perceive as unfairly harsh treatment at the hands of news organizations (Walk, 2020). In some instances, they have attributed this tendency towards hostility to their competition with the platforms for advertising dollars (Arnold, 2020; Walk, 2020).

In a study of Google News, researchers found that news stories about Google News tended to omit pertinent facts about Google News’ business model (e.g. that it does not accept advertising; Chyi et al., 2016). This study also found that news industry sources were by far the most prominent category of sources in stories about Google News. Ultimately, the authors concluded that ‘Google was portrayed as a foe of the news industry, with two-thirds of the articles portraying the Google–newspaper relationship as competitive or hostile’ (Chyi et al., 2016: 807).

And so, while the previous section raised questions about whether the dependencies inherent in the relationship between digital platforms and news organizations might discourage the production of critical platform beat reporting, the concern raised here is about the competitive relationship between platforms and news organizations potentially pushing platform beat reporting in the opposite direction, towards unjustly critical and negative reporting.

Technological complexity or opacity

All institutions are, to some extent, complex and opaque in ways that can affect news organizations’ ability to effectively perform their watchdog functions. Obtaining information from government agencies or corporations can often be difficult, despite various legal protections that facilitate journalistic access to information.

There are, however, additional layers of complexity and opacity inherent in digital platforms that make the platform beat particularly challenging for news organizations. First, there is the technical complexity that is at the core of the operation of these platforms, which results in a fairly high bar in terms of journalists with the necessary technical expertise to report on these platforms accurately and effectively. In addition, many of the technical details related to how these platforms operate remain within the proverbial ‘black box’ which inherently limits news organizations’ ability to conduct robust watchdog reporting. Finally, there is the fact that many of the data-intensive approaches to investigative journalism that characterize the platform beat require types of reporting and analytical skills that are not yet widespread in the journalistic profession. The result of this combination of factors is a potentially substantial knowledge gap between news organizations and the platforms that they cover.

This imbalance can potentially put news organizations in a somewhat subservient position relative to the platforms. For instance, in a study of mainstream news outlets’ reporting on artificial intelligence, Brennen et al. (2018) found that such coverage was largely driven by industry products, initiatives, or announcements, and tended to lack a critical dimension. Such findings suggest that news organization may operate to some extent as passive conduits for the platforms they cover, even serving more as cheerleaders for the platforms than critical watchdogs (Naughton, 2019).

This knowledge gap may also affect platform beat reporting in the opposite direction. Today, many argue that we are in the midst of a ‘techlash’, in which news organizations have swung radically in the other direction, and are engaging in excessively and unjustly negative coverage of these platforms (Gallup, 2020). Of course, any techlash-related critiques also need to address the fact that, as with any other institution, as digital platforms have become larger and more powerful, they are more in need of intensive journalistic scrutiny (Alex, 2020). It also important that we acknowledge the possibility that the platforms’ actions (and inactions) have merited what may in fact be an accumulation of overwhelmingly negative coverage.

That being said, tech sector critics have pointed to inadequate technical knowledge of some journalists on the platform beat as a potential contributing factor to excessively negative coverage (Kantrowitz, 2020). Essentially, reporters’ limited technical knowledge could lead them to instinctively gravitate to negative angles on stories as a result of an inadequate understanding of all of the relevant technical considerations. Further pressures in this direction could be brought to bear from the competitive dynamics described earlier and from the economic dynamics discussed in the next section.

Any knowledge gap on this front is also a function of the fact that these digital platforms tend not to be forthcoming about the technical details of their operation, out of both competitive concerns and legitimate concerns about providing information that facilitates manipulation of the platforms by bad actors (Napoli, 2019). This is why Diakopoulos (2013) emphasizes the importance of digital watchdogs reverse-engineering the algorithmic decision-making systems that are at the core of these platforms. Of course, reverse-engineering algorithmic systems is a complex task, requiring knowledge and skills sets that are not part of the traditional journalist’s toolkit. As Diakopoulos (2013) notes, ‘Developing the human resources to do algorithmic-accountability reporting will take dedicated efforts to teach the computational thinking, programming, and technical skills needed to make sense of algorithmic decisions’ (p. 26).

These technical complexities associated with digital platforms can potentially undermine the rigour with which news organizations cover the platform beat; and as this section has illustrated, these complexities have the potential to bias platform beat coverage in either positive or negative directions.

Diminished economic resources

The kind of reporting described earlier is clearly resource-intensive, requiring substantial investments in time and computing resources, as well as in the construction of reporting teams with diverse skill sets (Elmer, 2020). In the contemporary media environment, in which the economic model of commercial journalism is suffering a continued and well-publicized implosion, this kind of resource-intensive reporting is particularly difficult for news organizations to conduct; or, at least will likely remain confined to a select few well-resourced news organizations. As research by Brennen et al. (2018) notes,

Specialty reporting – including science and technology journalism – has been especially impacted. Some outlets have reduced or even eliminated their science and/or technology desks. These changes mean that some outlets cover these stories less frequently, task non-specialist reporters with reporting these stories, give their reporters less time and fewer resources to cover them, or encourage more reliance on press releases or wire articles. (p. 2)

These declines in resources clearly strike at the core of the type of reporting that is necessary for news organizations to effectively cover the platform beat.

As should be clear, then, the platform beat is a uniquely complex and labour intensive beat to cover, yet most news organizations are on a steady path to being less – rather than more – capable of investing the resources necessary to address the challenges posed by this beat. And so, as much vital digital platform reporting as we have seen in recent years, and as impactful as much of this reporting has proven to be, there may very well exist a but-for world in which more news organizations are covering the platform beat far more rigorously and extensively than we are currently witnessing.

In addition, according to some in the tech sector, news organizations may be incentivized to provide negative coverage of digital platforms as a result of the pressures emanating from their current financial situation. Arnold (2020), for instance, contends that negative, techlash-type headlines are good drivers of clicks; and that, in the challenging and competitive environment in which news organizations operate today, journalists are operating under increasing pressures to publish quickly, which can lead to inaccuracies.

Whether digital platforms are receiving too much, too little, or just the right amount of critical coverage in light of how they are operating and being used is beyond the scope of this analysis. The point here has been to consider how the declining financial status of many news organizations may impact how well these organizations perform on the platform beat.

Discussion

Digital platforms have emerged as powerful institutions in the media ecosystem and in the broader dynamics of the relationships between citizens and other governmental and economic institutions. Yet, evidence to date suggests that these platforms lack the ability and/or the incentive to fully and comprehensively police themselves for abuses and manipulations by bad actors (Napoli, 2019). In some instances, the platforms themselves have proven to be bad actors, engaging in deceptive practices or violating users’ trust in the treatment of their data (Wiley, 2019). However, the prospect of more intensive government oversight in this space raises long-standing and well-founded concerns about the dangers of government intervention in the realms of speech and journalism. In many national contexts, even the prospect of a moderate governmental oversight role is effectively thwarted by the scope of the legal doctrines that prevent government intrusions upon individual speech rights (Napoli, 2019).

This state of affairs leaves news organizations as perhaps the most viable and important line of defence in terms of preventing the wide range of social and political harms that can arise from the misuse and ineffective governance of digital platforms. Such reporting, it should be emphasized, need not lead to governmental actions of any kind in order to have an impact.³ Research has shown that the public awareness generated by such reporting is capable of leading to important changes in platform behaviour (Barrett and Kreiss, 2019; The Markup, 2021). This is the essence of the watchdog function of journalism and the kind of tangible impact that it can have.

In addition, the platform beat should be considered in relation to what Anderson et al. (2014) have called the scarecrow function of journalism. Under the scarecrow model, the very presence of an attuned and vigorous press deters institutional actors from engaging in at least some harmful activities. In this regard, the value of a robust platform beat is not fully captured by the harmful activities that news organizations uncover. It is also captured by the invisible deterrent effect it has had on additional harmful actions.

And yet, as this article has illustrated, news organizations face a range of challenges to effectively covering the platform beat. Some of these challenges arise from the institutional interdependencies in the relationship between digital platforms and news organizations. Others arise from the difficulties inherent in effectively covering digital platforms, and from the current financial challenges facing news organizations.

These dynamics that have been described here put platforms in a uniquely powerful position, compared to other institutions, to affect how effectively news organizations cover them, and even to affect the extent to which such reporting circulates in the news and information ecosystem. This situation bears directly on contemporary concerns about institutional trust, as the platforms’ technical–industrial model of trust gets layered upon – and ultimately impacts the efficacy of – the institutional–professional model of trust that journalistic institutions are working to apply to how the platforms are performing in the application of their still-evolving technical–industrial trust model.

This situation points to a need to disentangle the relationship between platforms and the press (Napoli, 2019). We have seen some movement in this regard, with some news organizations reducing their dependence on platforms for distribution (Meese and Hurcombe, 2020). We are beginning to see critics raise more questions about the implications of the sizable financial contributions that some platforms are providing to news organizations (Usher, 2020). And, perhaps most important, we are seeing the rise of independent, specialized news organizations that focus exclusively on digital platforms and are staffed with journalists possessing the necessary knowledge and skill sets to do so effectively (e.g. Angwin, 2020).

To the extent that the future of journalism and of informed citizenries seem to be inextricably intertwined with the governance and trust models employed by large digital platforms, the platform beat may very well be the most uniquely complex, challenging and consequential institutional beat in contemporary journalism. The complexities and tensions laid out in this article are intended to provide baseline propositions for future empirical work exploring the dynamics of this beat, as well as justifications for enhanced institutional support for the organizations and individuals serving as vital watchdogs of contemporary digital platforms.

Footnotes

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: This research was made possible in part by grants from the Carnegie Corporation of New York and the Knight Foundation. The statements made and views expressed are solely the responsibility of the author.

Notes

References

Alex (2020) Tech, media, and what journalism is for. Substack, 15 February. Available at: https://alexwrites.substack.com/p/tech-media-and-what-journalism-is.

Anderson

Bell

Shirky

(2014) Post-industrial journalism: Adapting to the present. Available at: https://academiccommons.columbia.edu/doi/10.7916/D8N01JS7.

Angwin

(2020) Are journalists and researchers the real content moderators? The Markup, 19 September. Available at: https://www.getrevue.co/profile/themarkup/issues/are-journalists-and-researchers-the-real-content-moderators-277640.

Arnold

(2020) What is the July Twitter spat between Taylor Lorenz and Balaji S. Srinivasan really about? Quora, 12 July. Available at: https://www.quora.com/What-is-the-July-2020-Twitter-spat-between-Taylor-Lorenz-and-Balaji-S-Srinivasan-really-about.

Baker

(1994) Advertising and a Democratic Press. Princeton, NJ: Princeton University Press.

Barrett

Kreiss

(2019) Platform transience: Changes in Facebook’s policies, procedures, and affordances in global electoral politics. Internet Policy Review 8(4): 1–22.

Benton

(2020) Google is giving $1 billion to news publishers – To help convince governments not to take a whole lot more than that. Nieman Lab, 1 October. Available at: https://www.niemanlab.org/2020/10/google-is-giving-1-billion-to-news-publishers-to-help-convince-governments-not-to-take-a-whole-lot-more-than-that/

Blecher

(2019) Digital News Innovation Fund: Three years in and 662 total projects supported. Google in Europe, 21 March. Available at: https://www.blog.google/around-the-globe/google-europe/digital-news-innovation-fund-three-years-and-662-total-projects-supported/

Brennen

Howard

Nielsen

(2018) An industry led debate: How UK media cover artificial intelligence. Fact Sheet, December, Reuters Institute for the Study of Journalism, Oxford.

10.

Briant

(2021) The grim consequences of a misleading study on misinformation. Wired, 18 February. Available at: https://www.wired.com/story/opinion-the-grim-consequences-of-a-misleading-study-on-disinformation/

11.

Brown

(2020) Facebook invests additional $100 million to support journalism during the coronavirus crisis. Facebook Journalism Project, 30 March. Available at: https://www.facebook.com/journalismproject/coronavirus-update-news-industry-support.

12.

Calzada

Gil

(2020) What do news aggregators do? Lessons from Google News in Spain and Germany. Marketing Science 39(1): 134–167.

13.

Caplan

boyd

(2018) Isomorphism through algorithms: Institutional dependency in the case of Facebook. Big Data & Society 5(1): 1–12.

14.

Chyi

Lewis

Zheng

(2016) Parasite or partner? Coverage of Google News in an era of news aggregation. Journalism & Mass Communication Quarterly 93(4): 789–815.

15.

Coddington

(2019) Aggregating the News: Second-Hand Knowledge and the Erosion of Journalistic Authority. New York: Columbia University Press.

16.

Cowan

Westphal

(2010) Public policy and funding the news. Center on Communication and Leadership Policy Research Series, January. Available at: https://www.niemanlab.org/pdfs/USC%20Report.pdf.

17.

Diakopoulos

(2013) Algorithmic accountability reporting: On the investigation of black boxes. Report from the Tow Center for Digital Journalism. Available at: https://www.cjr.org/tow_center_reports/algorithmic_accountability_on_the_investigation_of_black_boxes.php.

18.

Diakopoulos

(2019) The algorithms beat: Angles and methods for investigation. The Data Journalism Handbook, 27 September. Available at: https://datajournalismhandbook.net/handbook/two/investigating-data-platforms-and-algorithms/the-algorithms-beat-angles-and-methods-for-investigation.

19.

Editorial Board (2018) The poison on Facebook and Twitter is still spreading. New York Times, 19 October. Available at: https://www.nytimes.com/2018/10/19/opinion/facebook-twitter-journalism-misinformation.html.

20.

Elmer

(2020) Algorithms in the spotlight: Collaborative investigations at Spiegel online. In: Gray

Bounegru

(eds) The Data Journalism Handbook. Amsterdam: Amsterdam University Press. Available at: https://datajournalism.com/read/handbook/two/investigating-data-platforms-and-algorithms/algorithms-in-the-spotlight-collaborative-investigations-at-spiegel-online.

21.

Facebook (n.d.) Inauthentic behavior. Facebook community standards. Available at: https://www.facebook.com/communitystandards/inauthentic_behavior.

22.

Fischer

(2020) Media’s failed attempt to take on the Facebook-Google ‘duopoly’. Axios, 29 September. Available at: https://www.axios.com/medias-failed-attempt-to-take-on-the-facebook-google-duopoly-f6011c72-0209-4e49-9f6d-5c52b9212245.html.

23.

Gallup (2020) Techlash? America’s growing concern with major tech companies, March. Available at: https://knightfoundation.org/wp-content/uploads/2020/03/Gallup-Knight-Report-Techlash-Americas-Growing-Concern-with-Major-Tech-Companies-Final.pdf.

24.

Gandy

Jr (1982) Beyond Agenda-Setting: Information Subsidies and Public Policy. Westport, CT: Praeger.

25.

Grygiel

(2019) With cryptocurrency launch, Facebook sets it path toward becoming an independent nation. The Conversation, 19 June. Available at: https://www.pri.org/stories/2019-06-19/cryptocurrency-launch-facebook-sets-its-path-toward-becoming-independent-nation.

26.

Hutchinson

(2020) Google announces $1 billion partnership program with news publishers in Europe. Social Media Today, 1 October. Available at: https://www.socialmediatoday.com/news/google-announces-1-billion-partnership-program-with-news-publishers-in-eur/586281/

27.

Ingram

(2010) For better or worse, Google is a nation-state. Gigaom, 11 August. Available at: https://gigaom.com/2010/08/11/for-better-or-worse-google-is-a-nation-state/

28.

Ingram

(2018) The platform patrons: How Facebook and Google became two of the biggest funders of journalism in the world. Columbia Journalism Review, 16 May. Available at: https://www.cjr.org/special_report/google-facebook-journalism.php.

29.

Kafka

(2019a) Rupert Murdoch wanted Mark Zuckerberg to pay him for news stories – and now Facebook is going to do just that. Vox, 24 October. Available at: https://www.vox.com/recode/2019/10/24/20929919/facebook-zuckerberg-murdoch-news-publishers-pay-content.

30.

Kafka

(2019b) The logic behind Apple’s give-us-half-your-revenue pitch to news publishers. Vox, 13 February. Available at: https://www.vox.com/2019/2/13/18224013/apple-news-publishers-50-percent-subscription-explained-new-york-times-washington-post-texture.

31.

Kantrowitz

(2020) Why people can’t stand tech journalists: An interview with Casey Newton. Onezero, 21 August. Available at: https://onezero.medium.com/why-people-cant-stand-tech-journalists-an-interview-with-casey-newton-b5ab0826bc2a.

32.

Kirkpatrick

(2018) The Facebook defect. Time, 12 April. Available at: https://time.com/5237458/the-facebook-defect/

33.

Kosslyn

(2017) Fact-check now available in Google Search and News around the world. Google, 7 April. Available at: https://blog.google/products/search/fact-check-now-available-google-search-and-news-around-world/

34.

LaForme

(2018) The next big thing in journalism might be algorithm reporters. Poynter, 15 March. Available at: https://www.poynter.org/tech-tools/2018/the-next-big-thing-in-journalism-might-be-algorithm-reporters/

35.

Lunden

(2020) Twitter is donating $1 million across two foundations to support journalism during the coronavirus pandemic. TechCrunch, 24 March. Available at: https://techcrunch.com/2020/03/24/twitter-is-donating-1m-across-two-foundations-to-support-journalism-during-the-coronavirus-pandemic/

36.

Lytvynenko

(2020) Here’s a running list of false and misleading claims about Trump’s coronavirus diagnosis. BuzzFeed, 5 October. Available at: https://www.buzzfeednews.com/article/janelytvynenko/false-and-misleading-information-trump-coronavirus.

37.

McGregor

(2020) What even is ‘coordinated inauthentic behavior’ on platforms? Wired, 17 September. Available at: https://www.wired.com/story/what-even-is-coordinated-inauthentic-behavior-on-platforms/

38.

Meese

Hurcombe

(2020) Facebook, news media, and platform dependency: The institutional impacts of news distribution on social platforms. New Media & Society. Epub ahead of print 3 June. DOI: 10.1177/1461444820926472.

39.

Mozur

(2018) A genocide incited on Facebook, with posts from Myanmar’s military. New York Times, 15 October. Available at: https://www.nytimes.com/2018/10/15/technology/myanmar-facebook-genocide.html.

40.

Napoli

(2014) Automated media: An institutional theory perspective on algorithmic media production and consumption. Communication Theory 24(3): 340–360.

41.

Napoli

(2019) Social Media and the Public Interest: Media Regulation in the Disinformation Age. New York: Columbia University Press.

42.

Napoli

Caplan

(2017) Why media companies insist they’re not media companies, why they’re wrong, and why it matters. First Monday 22(1). Available at: https://firstmonday.org/ojs/index.php/fm/article/view/7051.

43.

Napoli

Royal

(in press) Platforms and the press: Regulatory interventions to address an imbalance of power. In: Flew

Martin

Gillett

(eds) Digital Platform Regulation: Beyond Transparency and Openness. London: Palgrave Macmillan.

44.

Naughton

(2019) Don’t believe the hype: The media are unwittingly selling us an AI fantasy. The Guardian, 13 January. Available at: https://www.theguardian.com/commentisfree/2019/jan/13/dont-believe-the-hype-media-are-selling-us-an-ai-fantasy.

45.

News Media Alliance (2019) Google benefit from news content, June. Available at: http://www.newsmediaalliance.org/wp-content/uploads/2019/06/Google-Benefit-from-News-Content.pdf.

46.

Owen

(2019) Pro Publica’s Facebook monitoring political ad tool (which Facebook fought) is alive again with a new home at the Globe and Mail. Nieman Lab, 7 June. Available at: https://www.niemanlab.org/2019/06/propublicas-facebook-monitoring-political-ad-tool-which-facebook-fought-is-alive-again-with-a-new-home-at-the-globe-and-mail/

47.

Paul

(2020) Here are all the steps social media made to combat misinformation. Will it be enough? The Guardian, 30 October. Available at: https://www.theguardian.com/technology/2020/oct/29/here-are-all-the-steps-social-media-made-to-combat-misinformation-will-it-be-enough.

48.

Tardáguila

Benevenuto

Ortellado

(2018) Fake news is poisoning Brazilian politics. WhatsApp can stop it. New York Times, 17 October. Available at: https://www.nytimes.com/2018/10/17/opinion/brazil-election-fake-news-whatsapp.html.

49.

The Markup (2021) In the days following our initial report, Facebook’s recommendations for political groups plummeted. Twitter, 14 February. Available at: https://twitter.com/themarkup/status/1361121548501848064.

50.

Tracy

(2019) Craig Newmark, new friend to journalism, gives $6 million to consumer reports. New York Times, 6 June. Available at: https://www.nytimes.com/2019/06/06/business/media/consumer-reports-craig-newmark.html.

51.

Usher

(2020) Beggars and choosers: How Google and Facebook compromise media independence with their corporate donations. Open Markets Institute, 30 September. Available at: https://www.openmarketsinstitute.org/publications/beggars-and-choosers-facebook-google.

52.

van Dijck

(in press) Governing trust in European platform societies: Introduction to the EJC special issue. European Journal of Communication.

53.

Wakabayashi

(2020) Election misinformation continues staying up on YouTube. New York Times, 10 November. Available at: https://www.nytimes.com/live/2020/2020-election-misinformation-distortions.

54.

Walk

(2020) What’s wrong with tech folks who attack the tech media. And what’s wrong with tech media today, 15 February. Available at: https://hunterwalk.com/2020/02/15/whats-wrong-with-tech-folks-who-attack-the-tech-media-and-whats-wrong-with-tech-media-today.

55.

Wiley

(2019) Mindf*ck: Cambridge Analytica and the Plot to Break America. New York: Random House.

56.

Wilson

(2020) Australia has 17 million Facebook users and seven Facebook fact checkers. BuzzFeed, 19 February. Available at: https://www.buzzfeed.com/cameronwilson/australia-facebook-seven-fact-checkers-17-million-users.