Proposal Description

Trust in official statistics is key to evidence-based policymaking, public engagement and democratic accountability. Yet this trust is increasingly tested by misinformation, declining confidence in institutions, fragmented information ecosystems and the rapid development of AI-driven data production and dissemination. This session will examine how trust in official statistics is measured, what drives or undermines it, and how statistical institutions can respond through quality, transparency, relevance, responsible innovation and stronger engagement with users and citizens. Recent research emphasises that trustworthiness is built through quality, value, openness, accessibility, clear communication, and honest explanation of methods, uncertainty and limitations (Rodgers, 2026). However, trust is no longer secured by technical quality alone. Statistical institutions must also remain relevant, independent, transparent and understandable in an ever-changing information environment. National statistical offices face pressures on both competence and integrity, in particular when official figures appear disconnected from people’s lived experience, or when political polarisation and digital information flows make impartial evidence harder to recognise (Gennari, 2026). AI adds further complexity: while it may support statistical production, quality management and dissemination, it also raises risks linked to hallucinated or misleading outputs, bias, confidentiality, governance and reputational damage if users cannot understand or trust how AI-assisted statistics are produced (MacFeely, 2026). This session approaches trust as a multidimensional challenge. Contributions may examine recent evidence and emerging methodologies for measuring public trust in official statistics, along with related challenges, including AI-generated disinformation, algorithmic mediation, data privacy, public understanding of statistics, the communication of uncertainty, and the gap between official indicators and lived experience. The session will encourage dialogue between official statisticians, researchers, policymakers, communicators, regulators and international organisations on how trust in official statistics can be strengthened through high-quality statistical production, transparent methods, responsible use of AI and new data sources, effective communication, citizen engagement, statistical literacy and safeguards for independence, privacy and confidentiality.

Rodgers, H. (2026). Trust and official statistics.

Gennari, Pietro. "Can official statistics survive the collapse of public trust?." Statistical Journal of the IAOS 42.1 (2026): 3-10.

MacFeely, Steve. "The role of AI in Official Statistics: Black hole or worm hole?." Statistical Journal of the IAOS (2026): 18747655251410711.