Answers are grounded in this paper and cite their sources.
As a sociotechnical practice, fact-checking integrates artificial intelligence (AI) while resisting “enchanted determinism” to preserve professional agency and epistemic authority. This study examines how fact-checkers in Europe and Latin America perceive AI affordances within their professional practices and how organizations articulate and govern AI use. Methodologically, the study is based on 50 qualitative interviews with practitioners and a thematic analysis of 18 publicly available AI guidelines and institutional documents. The findings suggest a distinction between instrumental optimism and epistemic disenchantment . While fact-checkers embrace AI’s functional affordances to enhance efficiency in tasks such as monitoring and data processing, they resist delegating the authority to determine truth to autonomous systems. At the organizational level, human-in-the-loop requirements formalize this distinction and operate as a form of boundary work that defines the limits of automation. We also identify a shared rationale for AI use across countries, which may reflect some degree of institutional convergence. However, this consensus coexists with a persistent “AI divide” shaped by contextual forces. Overall, the study shows how fact-checkers act as sociotechnical agents who “strip the magic” from AI, reaffirming human judgment as the field’s core epistemic foundation.
No takes yet. Share an insight, caveat, or question.
Cazzamatta et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: