“When politicians make claims, we are journalists”: Political fact-checking and social media partnerships

---

Implications 

Emerging from journalistic traditions of political accountability, the fact-checking movement was initially concerned with evaluating claims made by politicians and public officials seeking to shape public debates to hold the behaviour of those in power to account. Over the past decade, however, the field has been shifting towards the verification of rumours spread by anonymous actors online, in part due to the shrinking journalism sector and the growing influence of online media (Amazeen, 2017; Cazzamatta, 2024; Lelo, 2023). This “debunking turn” arguably sees fact checkers pivot away from their original mission of political fact-checking, a shift associated with the 2016 launch of Meta’s 3PFC program (Graves et al., 2023) and the subsequent launch of similar projects by ByteDance (TikTok) and other partnership strategies by YouTube/Google and X. The effects of these programs, and Meta’s in particular,¹Our focus in this article is on Meta for several reasons, including scale, the outspokenness of Meta CEO Mark Zuckerberg in suggesting that third-party fact-checking has been central to the company’s strategy for content moderation (which was later rolled back at the beginning of the second Trump administration in 2025), and the more formal aspects of contracted partnerships. have been significant. These programs are often credited with the global proliferation of fact-checking organizations (Bélair-Gagnon et al., 2023) and for broadening the umbrella of who, and which organisations, are doing this work. 

Over 100 fact-checking organizations across 115 countries have partnered with Meta to address viral misinformation online under Meta’s 3PFC umbrella (Catalanello & Sanders, 2025). Essentially building fact checkers into the online content moderation system, the partnership was viewed by many fact checkers as an opportunity to amplify their work in mitigating information disorder online (Bélair-Gagnon et al., 2023). Our interviewees cited Meta as a vital funding source for their fact-checking work, despite certain limitations, and the research took place in the context of significant uncertainty around the sustainability of the sector in Europe, given, in part, the termination of Meta’s 3PFC program in the United States at the beginning of 2025. Studying fact checkers’ negotiations of their journalistic missions and the terms of their platform partnerships is particularly critical as public concern grows over Big Tech’s exercise of funding strategies to “capture” journalism and growing criticism of fact checkers (Papaevangelou, 2023).

Whether through their community standards, support for umbrella organizations, such as the International Fact-Checking Network (IFCN) at the Poynter Institute or the European Fact-Checking Network Standards (EFCSN), or the types of facts that they pay fact checkers to check, tech companies have been involved in writing the rules, including minimum standards to abide by and what type of organization can legitimately “do fact-checking.” While many organizations have emerged outside of this network, with different models of funding or goals, the ties between Big Tech and fact-checking in the EU are particularly influential. The exclusion of politicians’ own claims, posts, and ads being ineligible for fact-checking within Meta’s 3PFC program has prompted concerns that the program may have significantly altered the nature of fact checkers’ work, which initially grew out of the media and journalism sector with a focus on holding the speech of politicians to account (Mantas, 2021). Prior to our study, there has been little research on how such cross-sector partnerships affect European fact checkers’ understanding of their professional mission and norms (for exceptions, see the research by Bélair-Gagnon et al., 2023; and Farkas & Bengtsson, 2026).

Our research starts from a tension that can be easily overlooked: fact checkers largely see themselves as journalists—their occupational ideology is closely tied to journalistic pillars of fairness, objectivity, and accuracy (Graves & Cherubini, 2016)—yet some platform partnerships restrict some activities that are central to defining that identity. When 3PFC imposes a blanket prohibition of assessing the accuracy of politicians’ speech, they risk splitting fact checkers’ role conception from role performance (Mellado & Van Dalen, 2013). Our puzzle, then, is not simply whether platforms shape fact-checking but how fact checkers respond when the rules of the partnership contradict the norms and expectations of the profession. We examine European fact checkers’ experiences of platform partnerships, focusing analytically on Meta’s 3PFC programme because interviewees described it as a benchmark model for the sector. Identifying how fact checkers perceive and respond to platform policies and logics within the 3PFC programme, we outline the strategies that interviewed fact checkers report adopting to minimise the gap between how they conceptualise their roles and how theyperform these in practice.

Rather than treating the platforms as the sole architects of current practice, or fact checkers as simply implementers, we examine the dynamic interplay between partnership structures and fact checkers’ agency (Rifqi, 2025). This is an increasingly urgent issue with significant structural questions: if platforms are significantly (or entirely) reducing support for professional fact-checking while regulators are simultaneously demanding evidence of risk mitigation, what are the implications for political accountability? 

Partnership structure and fact checker’s agency: A dynamic interplay 

Previous literature has examined the normative trade-offs that cross-sector partnerships might produce in fact checkers’ working methods (Bélair-Gagnon et al., 2023) or whether they might move fact checkers away from their original mission (Graves et al., 2023). But research has generally neglected to examine the strategies that fact checkers describe adopting to bridge tensions that emerge between their role conception and platform rules around fact-checking politicians’ speech. Studies that have gauged how political fact-checking is influenced by external constraints have tended to focus on authoritarian contexts outside of the EU (Badji et al., 2024; Fakida, 2021; Liu & Zhou, 2022). Research has also tended to rely mainly on quantitative methods, tracking “number of political fact-checks published” (Fakida, 2021) and excluding fact checkers’ own reflections on their mission, barriers to their political fact-checking work, and strategies they have adopted to resist these.

Through qualitative interviews and interrogating the strategies that fact checkers describe adopting in response to tensions with platform rules, we explore a dynamic interplay between structure and agency in the context of very large online platforms (VLOPs) fact-checking partnerships. We find evidence that fact checkers in partnership with Meta insist on fact-checking politicians outside the partnership and have pushed for more transparency on Meta’s rules on political speech. We note how while social media partnerships like the 3PFC programme provide fact checkers with rules, they simultaneously permit the creation of “new rules” through repeated action and interaction, thereby presenting a dynamic outlook on the dialectical relationship between journalistic role conceptions and changes in conditions of work (Grubenmann & Meckel, 2015)—in essence, a form of negotiated governance. 

These findings point to how fact checkers might be better empowered, or disempowered, within these structures, as is directly relevant to current EU regulation, including the DSA and the Code of Conduct on Disinformation, which includes a commitment for cooperation between online platforms and fact checkers (Botan et al., 2025). Moreover, at a time when political actors increasingly use VLOPs to communicate with European voters, our research provides important context into fact checkers’ ongoing demands for further discussions, including at policy level, on how Meta’s treatment of speech by politicians impacts democratic debate (Bickerton and Acetti, 2018; Full Fact, 2020).

Evidence 

Finding 1: Fact checkers identify as journalists when fact-checking politicians’ speech. 

Without negating the heterogeneity of the over 400 organizations in the global fact-checking movement, our data indicates that interviewed fact checkers generally identify with journalistic values, objectives, and standards, and that this identification endures even in an age of social media partnerships (Cazzamatta et al., 2024). This supports earlier studies that point to the journalistic institutional roots of the global fact-checking movement (Graves & Cherubini, 2016)

Nearly 80% of EU survey respondents agreed that fact checkers in their organization were best described as journalists, as compared to civil society (13%), academics (4.5%), and activists (0%). Rather than a statistical representation of the whole European sector, this finding lends context to an overarching trend in our interviewees’ alignment with journalistic values. Interviewee 5, for example, claimed that “the vast majority of the European fact checkers would describe themselves as journalists,” albeit a niche area as described by another interviewee:

[Fact checkers] may be a specialised kind of journalist. I consider us a kind of sub-branch of investigative journalism. If we look at it across disciplines, that’s what we do—we investigate, and we inform the public. Beyond that, we also have other roles. That’s why I say we’re a broader organization. We do journalism, yes, but our role is wider. (Interviewee 20)

Conversely, Interviewee 20 also noted how journalistic identity can have the opposite effect, in countries where “people’s perception [is] that being a journalist means corruption.” Even then, the interviewee affirmed that fact checkers in their organization are journalists “in the broader sense.”

Journalism, as a framework and occupational ideology, allows fact checkers to give meaning and legitimacy to their work, establishing their position in society and vis-à-vis other professions (Deuze, 2005). Alluding to journalistic pillars like objectivity can potentially protect fact checkers from risks of the trade, such as public criticism, particularly as fact checkers cite public trust to be an enduring challenge (Interviewee 2; Tuchman, 1972). Interviewee 17, for instance, highlighted the need to set aside personal convictions when verifying claims.

Fact checkers link journalistic principles, such as removing “bias” and “personal values,” to their approach in fact-checking elite political speech. The commitment to continuing to fact-check political elites, despite the exclusion from Meta’s 3PFC program, is linked to fact checker’s journalistic role-conception, or “self-image” (Interviewee 18). 

I think, from doing journalism, you kind of have to leave, to some extent, your personal values at the door. You’ve got to be aware of your own biases … That may involve, say, fact-checking of politicians … where you go into a fact check, and you think, oh, that person is wrong, or I don’t necessarily agree with this person … (Interviewee 6)

One of Meta’s stipulations is that claims made by politician accounts … are not allowed to be checked … arguing that politicians are constantly subject to a kind of public fact check anyway due to their public profile. You can agree with that or not. We are, of course, not bound by it in our work insofar as … we are journalists. Of course, our organization checks whether [the politicians’ claims] are true or not. That’s our core business. But that’s a rule that we have to abide by in our specific cooperation with Meta … but it doesn’t affect our self-image because we also work beyond that. (Interviewee 18) 

However, given the terms of these partnerships, fact checkers in social media partnerships necessarily fact-check online information beyond politicians’ speech. Interviewee 12 notes that, for their fact-checking organization, this has been an expansion, rather than a change in a fundamental goal of challenging those in power. 

We started doing fact-checking on a regular basis, like traditional fact-checking platforms like PolitiFact in the US or Full Fact in the UK. We covered mainly statements made by politicians because the idea behind it was not necessarily to have a journalistic program, but rather an accountability project that keeps politicians engaged with the promises that they made or trying to push them towards saying the truth when it comes to public policies …

With the war in Ukraine, something shifted in our approach. We decided to cover also other types of disinformation, so not only the statements made by politicians, but also social media narratives. (Interviewee 12)

Finding 2: Fact checkers in the 3PFC program indicate remaining active in fact-checking politicians.

Meta’s exclusion of political fact-checking from the 3PFC program policies cites a “fundamental belief in free expression” and “respect for the democratic process” (“Meta Programme Policies”). Interviewee 18 suggests the platform has justified this exclusion on the premise that politicians’ speech in democracies with a free press is already scrutinised. Many interviewees, however, disagreed with this justification; the rule against fact-checking politicians as part of the program appears to strike at fact checkers’ journalistic mission. 

The Meta/Facebook restrictions seem to be entirely arbitrary, and the limitations on what categories of content we can rate result in some of the most widespread misinformation—that spread by politicians and political candidates and aspirants—not being rated. (Interviewee 9)

Interviewee 12 further remarked that the rule led to small-reach users being fact-checked, while a politician can spread disinformation “without any possibility whatsoever of sanctioning from the platforms.” Despite this rule, our interviewees expressed remaining wedded to their “self-image” and journalistic roles (Interviewee 18). They insist on trying to overcome this “cash of coverage,” as stated by Interviewee 12, by fact-checking politicians on their platform. This might point to the possibility of fact checkers using funds from Meta to subsidise their fact-checking of politicians outside the partnership (Graves & Mantzarlis, 2020).

If something really interesting from a politician was going viral on a social network, on Facebook or Instagram … we’re going to verify it anyway. But we will verify it from our own sources. (Interviewee 4)

Within Meta, we are not allowed to label politicians. But for our fact-checking platform, we do rate them. But this norm means that we cannot have [the label on Meta]: “this was fact-checked by independent fact checkers.” But what we do in our work is entirely our business. (Interviewee 12)

Our survey further indicates that, despite Meta’s prohibition, some fact checkers do not consider that their organisation conducts a low frequency of political fact-checking: over 70% of the surveyed respondents who cited actively working within Meta’s 3PFC programme indicated that they considered that their organisation fact-checked politician speech often or very often. Given the small sample, we treated these results as contextual rather than inferential, but they support the interviews, where fact checkers insist that, in their view, their participation in 3PFC does not necessarily correspond to an abandonment of fact-checking politicians. Instead, interviews point to how fact checkers’ reflexively monitor how their activities might be negatively influenced by these partnerships and how to mitigate this.

Of course, we see that they [Meta] are trying to buy some positive PR in the field … We believe we are still independent, even though we have significant funding from platforms … But we are aware of the possibilities where it could [influence our decision-making in every-day work], and we’re trying to acknowledge them and to make it a point that it doesn’t. (Interviewee 1)

Meta suggests content that we could fact-check on Meta platforms, but it doesn’t decide whether we fact-check or not. That means that, first of all, we can reject what they suggest … And we can also choose directly what we want to verify within those Meta platforms—with the exception of statements made by politicians. (Interviewee 4)

Fact checkers argued they were aware of how social media partnerships might impact their role performance, particularly given the importance of Meta’s funding, which was often described by interviewees as a key funding source. This suggests a regularised relation of autonomy and dependence between fact checkers and Meta, with fact checkers’ agency and own norms in dialectical relationship to this and a form of infrastructure (Ananny, 2020; Giddens, 1984). Fact checkers remain rhetorically critical of Meta’s “PR” motivations, referencing how, by partnering with independent, accredited fact checkers, the platform also seeks to self-servingly benefit reputationally from partnering with independent, accredited journalists. These findings align with Farkas and Bengtsson’s (2026) discussion on how fact checkers stress a boundary between their journalistic values and platforms’ own motivations. Clearly, as public criticism of Big Tech partnerships increases, fact checkers that do remain in partnerships must find ways to protect their credibility and journalistic missions (Bengtsson et al., 2024).

Finding 3: Fact checkers demand transparent platform rules in line with their journalistic values. 

In defining which politicians cannot be fact-checked in its program, Meta now specifies that claims by “candidates running for office, current office holders—and, by extension, many of their senior appointees such as heads of government agencies—along with political parties and their leaders” are not eligible to be fact-checked (“Meta program policies”). While some respondents note being able to exercise discretion in determining who constitutes a politician under these rules, others have worked to improve consistency in fact checkers’ interpretation of these rules and have pushed for different platform rules. 

The line between who is or isn’t a politician … we have frequently discussed it with Meta representatives or other fact checkers to establish the line that is also in line with their guidelines … It is also concerning that we know the people we are talking to are not the decision makers. The people who are in the content moderation departments, even if they wanted to, I don’t believe they could change anything. But they are contact points. (Interviewee 1)

There’s a bit of a grey area with people where you can’t say: Is this a politician or is this not a politician? They sometimes have their own private Facebook profiles or pages … It can happen that we might have put [a fact-check] on there after all, but there have never actually been any requests from Meta for us to take something down or anything like that. (Interviewee 17)

As relevant to Interviewee 1’s statement, some fact-checking organisations, including Full Fact, have also reported that Meta’s policy around fact-checking politicians’ speech can create confusion for fact checkers in practice, especially during electoral periods, and that requests for clarification from the platform are not always met (Full Fact, 2020). These interviews further indicate how Meta defines and operationalises its rule against fact-checking politicians’ speech while not exactly clarifying how the platform understands this speech category; this coincides with Anany (2020)’s discussion on platforms’ power in categorising political speech, with negative implications for collective self-governance.

Meta/Facebook says that we cannot rate political figures for misinformation. In [my country] the bar to be a political figure is very low—just run for a local county council, for example, and you can lie your heart out! We put this to Meta/Facebook, and they gave us slightly different rules, for example, people who had never held office or were in a party that had never won a seat could still be rated, at least before an election was called. (Interviewee 8)

Despite sometimes achieving “slightly different rules” through these interactions, fact checkers noted that a barrier to more significant change was a lack of access to higher-level decision-makers at Meta. Notably, however, one fact checker recounts how some fact checkers have attempted to leverage their experience with Meta and the authoritative power of the 3PFC program to alter the rules of other fact-checking partnerships beyond political fact-checking and in line with their missions.

Because Meta’s model was the benchmark, we’d often say to [TikTok], “If Meta does it this way, why don’t you?” and they’d respond accordingly. Now, Meta’s behavior sends the opposite message: that companies can defy the law. Reports already suggest that TikTok is considering cancelling its own program too. Meta is signaling, on a broader geopolitical level, … that they can simply ignore European laws with impunity … Meta’s move sets a precedent with huge ripple effects—not only in our field but across society. (Interviewee 20)

While such strategies face uncertainty with Meta’s recent termination of its program in the United States, our findings indicate how, rather than simply being embedded passively within Meta’s programme, or social media partnerships more widely, fact checkers report continually reflecting on how they might resist these terms of partnerships and uphold their fact-checking missions. 

Methods 

Between April and July 2025, we conducted 20 semi-structured interviews with fact checkers actively working across different fact-checking organisations across 10 EU countries.²This included interviews from Slovakia (2), the Czech Republic (2), Italy (1), Ireland (4), Austria (2), Romania (1), Spain (2), Belgium (1), Germany (4), and Greece (1). Interviewees held a range of job titles in their organisations, including founders, directors, and heads of fact-checking departments. Interview participants were recruited from personal connections or through cold contact from fact-checking organizations in the Duke Reporters’ Lab database of active fact-checking organisations. They were provided with an information sheet about voluntary participation in the study and a preliminary interview questionnaire. The right to anonymity was offered to all sources. For consistency purposes, we have removed any identifier marks.

Participants were interviewed individually, either online or in person. Whenever possible, the interview was conducted in the interviewee’s preferred language. With participants’ consent, all interviews were recorded and transcribed. Transcripts were analyzed on NVivo using a thematic analysis approach, according to which the data was categorised according to emerging themes (Braun and Clarke 2006). Patterns from the coding procedure emerged and were considered against the RQs.

Between June and August of 2025, we also disseminated a survey to fact-checking organizations around the world and received 113 responses. This paper draws on the 23 survey responses specifically gathered from fact checkers from 9 EU countries. Given the small sample size, the survey is intended only to provide context to interview data in this article.³Survey responses were received from fact checkers in Spain (8), Portugal (2), Greece (2), Czech Republic (2), Croatia (1), Germany (2), Ireland (2), France (3), and Slovakia (1). Respondents indicated their position in their organization as primarily directors (8), editors (5), fact checkers (2), and founders (4). Coordinators of various fact-checking networks, including the European Fact-Checking Standards (EFCSN), shared the online survey with their members. Institutional research partners also shared the survey with their fact-checking contacts. Survey questions were designed using Qualtrics software to test the relevant research questions and consisted primarily of Likert-type items and multiple-choice questions, allowing for neutral (e.g., “Neither agree nor disagree”) choices, as well as options to skip questions, to improve validity. 

We acknowledge some limitations to our methods that we attempted to address. To mitigate potential selection biases emerging from sampling interviewees from personal connections, we also attempted recruitment through cold contact from fact-checking organizations listed as active in the Duke Reporters’ Lab database. While some organizations did not respond to initial contact, we secured at least five contacts without any previous personal connections. We also highlight that the interviews and the survey presented in this paper are part of a broader research project examining fact-checking more generally and investigating its intersections with EU regulations. While over half of interviewees discussed working with Meta’s program, some declined to detail their relationships with social media platforms, as they were curtailed by non-disclosure agreements. Because this article focuses on Meta’s political-speech eligibility rules vis-à-vis the role conception of fact checkers, our analysis draws especially on interviews with participants who have engaged with the 3PFC, while the broader interview and survey corpus provides contextual evidence about professional role conceptions and platform partnerships.