The future of fact-checking in the algorithmic society

The evolving landscape of fact-checking: From heroes to villains

Fact-checking is not a new phenomenon, as journalists have long been responsible for scrutinising public claims for accuracy. In the early 2010s, however, fact-checking became central to the social media ecosystem of trust and safety, including content moderation, and platforms started funding fact-checking initiatives, incentivising a rapid expansion of the field. New fact-checking organisations were funded, and initiatives such as the International Fact-Checking Network (IFCN) prominently started to design the rules and standards of the field. In the mid-2010s, fact checkers were considered a potential solution to a growing misinformation problem (Graves et al., 2024).

The promises that fact-checking seemed to offer for addressing the harms of mis- and disinformation, however, have changed significantly over the past few years. Fact-checking faces three different, yet interrelated crises: (i) a technological crisis, (ii) a legitimacy crisis, and (iii) a financial crisis. The expansion of generative AI has started to compete with human-led fact-checking. Given the significant cuts to funding for fact-checking organisations since early 2025, when the second Trump administration came into office, and the improved operational efficiencies of AI tools, there has been a turn towards using AI to scale and speed up fact-checking. Furthermore, as AI capabilities expand, users themselves have increasingly used AI to verify content. This leads to epistemological questions that are directly related to the legitimacy of fact-checking: what is truth, and who decides whether something is true or not? Can machines provide qualified accounts of reality (Dierickx et al., 2026)? If anyone can verify claims, even AI systems, why do we need fact checkers, and who determines which organisations enjoy legitimacy? Although organisations such as the IFCN exist to provide greater legitimacy to fact checkers, the rise of alternative networks (Global Fact-Checking Network [GFCN]) and regional networks (European Fact-Checking Standards Network [EFCSN]) has offered alternative approaches, challenging the IFCN’s hegemony. Furthermore, the legitimacy crisis is compounded by the ascension of populist movements and more libertarian approaches to internet regulation that call for unrestricted freedom of expression. In their view, fact checkers go from heroes capable of solving the misinformation problem to villains working as censors of free speech and curbing narratives. Responding to these views of fact checkers as censors, Meta has scaled down its partnership with fact-checking organisations, swapping fact checkers for community notes. This has led to a financial crisis, at least in the United States (Meta Business Help Center, n.d.). 

This special issue 

This special issue comes at a particularly timely moment, as we seek to examine the uneasy intersection of these crises. Acknowledging the complexities of the political, technological, and financial situation of fact checkers, we approach the challenges from an interdisciplinary perspective, seeking to bridge the gap between law, regulation, and practice. The articles emerge from a workshop held in Oxford in December 2025 that brought together academics, fact checkers, and members of civil society to debate and discuss this transitionary and precarious moment for the future of fact-checking. Together, the contributions in this issue elucidate, in theory and in practice, the challenges, threats, and future prospects of the profession, together with the working practices of fact checkers. 

The contributors examine fact checkers at different levels of analysis. While some papers focus on working practices, others consider the fact-checking phenomenon more broadly and its relationship with law and society. As such, the contributions can be divided into two main groups: those which explore fact-checking from the top down, assessing it in relation to society or social media platforms, and those which present a bottom-up approach, exploring the working practices of fact checkers. Even though the contributions might take one approach over the other, the conclusions offer mixed perspectives, which, taken together, present a nuanced and multifaceted account of the contemporary fact-checking landscape.

The special issue begins by assessing the phenomenon of fact-checking from a bottom-up, practical perspective. Ramos, Jiménez-Cruz, and Hernández-Escayola provide a Commentary from the practical perspective of fact checkers. Building on the real-life example of the Valencia floods in 2024, they demonstrate the shortcomings of traditional fact-checking tools in tackling conspiracy theories. They argue how, after tackling misinformation from the floods, one narrative about hidden deaths at a parking structure gained traction. Maldita.es demonstrated the use of “argument-check” as a tool to tackle this kind of conspiracy theory. “Argument-check” shifts the natural logic of fact-checking from the question “is something true or not” to the question “how can something be true?” In other words, argument-checking provides the public with critical questions that challenge the premises of conspiracy theories and seeks to optimize fact-checking interventions. 

Also assessing the working practices of the profession, Brautović, Grkeš Tošović, and John’s Research Note focuses on transparency and accountability of fact checkers. Their article evaluates EFCSN signatories’ corrections (visibility, placement, timing) and their discursive features (explanatory detail, attribution of error, acknowledgment of affected actors), investigating what these reveal about accountability and transparency of EFCSN signatories. Their findings suggest correction practices are more dependent on organizational culture rather than shared operational norms, as the practice is not systematically supervised by standard-setting bodies.

In his Commentary, Cunliffe-Jones, founder of Africa Check, delves into a model designed to help fact checkers select claims. His model is based on the assumption that some forms of mis- and disinformation can cause substantive effects or harms. These encompass objectively verifiable substantive injuries to the interests of an individual, a group, or the wider society. Building on this, the model distinguishes between claims that can generate these harms and those that cannot. Ultimately, he argues, fact checkers should prioritize claims that generate these substantive harms.

The remaining four contributions explore fact-checking from a broader perspective. Instead of focusing on the working practices of the profession, they shed light on questions of epistemology and on the intrinsically delicate relationship between fact checkers and social media platforms.

The relationship between fact checkers and social media platforms is explored in the Research Articles by Botan; González, Stremlau, and Mota; and Bozkurt, Mosleh, and Margetts. Botan’s research situates this relationship within the EU’s Code of Practice on Disinformation, asking how fact checkers and social media platforms cooperate and what the nature of their relationship is. Her findings demonstrate that fact checkers occupy a peripheral position within the disinformation ecosystem. Despite being cited by the Code of Practice, social media platforms remain the primary implementers of the Code. Platforms are the ones, for instance, who have access to the relevant data needed to tackle mis- and disinformation, whereas fact checkers’ own access to such data remains limited. The power asymmetry between the two actors, therefore, diminishes the influence of fact checkers. 

González, Stremlau, and Mota further explore the relationship between fact checkers and social media platforms. The paper departs from the idea that fact-checking initially grew out of the journalism and media sector with a focus on holding politicians accountable for their claims. However, the partnerships with social media platforms significantly changed the rules by which fact checkers operate. Meta’s rules for fact-checking, specifically, prohibit the verification of political speech. Through 20 interviews and a survey with fact checkers, the authors assess fact checkers’ perspectives of their role identity in light of such a prohibition and how fact checkers navigate Meta’s ban on political fact-checking.

Bozkurt, Mosleh, and Margaretts’ research article contributes to broader discussions on the legitimacy and technological crises of fact-checkers. Against the backdrop of platforms’ growing shift towards community notes and AI-assisted verification, the authors draw on interviews with 29 fact-checkers to assess practitioners’ views on these emerging tools. While respondents acknowledged that these tools might expand fact-checking capacity and that they contribute to increasing the speed and scale of content verification, they also revealed important concerns. In particular, fact-checkers highlighted that community notes have a significantly slow publication process and that AI tools present inherent biases and limited contextual awareness. Accordingly, the authors advocate for a multi-tiered verification model, which integrates community notes, AI-assisted tools, and professional fact checkers.

Finally, Asmolov situates his Commentary within the rise of digital authoritarianism and the expansion of States’ capacity to deploy digital innovation for regime resilience. His contribution concerns epistemic sovereignty vis-à-vis the appropriation of fact-checking resources by authoritarian regimes. He argues that this appropriation works in two ways: (i) the repurposing of crowdsourced fact-checking, when fact-checking emerges as participatory propaganda, and (ii) the repurposing of AI systems, when generative AI systems start to replicate state-sponsored narratives. In his view, the authoritarian appropriation of fact-checking narratives is compounded by the fact that ambiguity is often seen as a failure, rather than an unavoidable condition of political life. To tackle this issue and allow for the recognition of ambivalence not as a failure, Asmolov advocates repurposing AI systems as arbiters of truth. Instead of promoting the use of AI systems as prescriptive tools capable of determining the truth, AI systems could be seen as tools that “surface competing interpretations, contextualize claims, and make visible the assumptions underlying different narratives.” 

Together, the articles in this special issue demonstrate how fact-checking is being significantly transformed, both in intention and practice. The role fact checkers exhibit in judgement, contextual knowledge, and public accountability is central to their distinctive approach and is not easily reproduced or replicated by AI. Yet at the same time, AI will have a central role in re-writing and re-negotiating the rules and frameworks that fact-checking systems operate with in the future. Similar to other sectors, there are calls for greater transparency—particularly in terms of what data or models are being used to inform given judgements. This is particularly important from a global perspective, where issues of regional biases (with most of the data coming from larger markets in Europe and North America) arise from being used in markets that are turning more quickly to AI fact-checking due to significant resource limitations to employ human moderators at scale (such as many markets in the Global South). The future rules of fact-checking will increasingly be negotiated in a context where AI is very much already embedded in the practice, but how visible, transparent, and understood those terms are, and how clear they are to the public and profession they serve, remains to be seen.