Abstract

In contemporary information ecosystems dominated by cognitive overload, algorithmic fragmentation and declining trust in institutions, communicating official statistics has become structurally more challenging. We are increasingly overwhelmed by content: four in ten people intentionally avoid the news, and 31% do so due to information overload. In this context—echoing Stepfanie Tyler’s reflection—the work of informing (and being informed) is no longer only a matter of production but increasingly one of subtraction, cultivating what she calls a “new intelligence”: taste.
Recent scholarship describes this environment as the Age of Epistemia: a condition in which the formal coherence of AI-generated or algorithmically amplified content begins to replace empirically grounded knowledge, turning plausibility into a surrogate for truth (Quattrociocchi et al., 2025). For National Statistical Institutes (NSI), this raises a central question: how can statistical evidence remain recognisable, trustworthy and comprehensible in such an environment?

Theoretical framework
Research shows that trust in official statistics relies not only on methodological soundness but also on the recognisable human, professional and institutional origins of data (Allegrezza, 2018; Eurostat, 2022). In post-truth contexts, human presence acts as a cognitive authenticity cue: it increases credibility, supports expert-recognition mechanisms and enhances the processing of complex information (Sundar & Nass, 2000; Metzger & Flanagin, 2013).

Implementation and innovation
This paper presents Istat’s strategic response: a systematic shift toward human-centred, researcher-led video communication supported by generative AI tools. These tools enable rapid transformation of technical statistical language into short explanatory scripts, allowing timely video production aligned with each statistical release. Within a landscape marked by infodemia and datademia, the project demonstrates how humanising dissemination can reinforce the authority and recognisability of official statistics.
Despite a −90.9% reduction in the total number of posts, overall communication performance improved significantly. Cross-media impressions reached 6.38 million (+59%), follower growth amounted to +8.6%, and the overall engagement rate increased from 2.75% to 4.43% (+61.2%). The most substantial change concerns video dissemination: video views increased by more than +1,100%, surpassing 2.33 million, while YouTube views rose from 133,835 in 2024 to 2,284,037 (+1,600%) in 2025.
The Statistics in video series—featuring researchers directly presenting statistical releases—emerged as the strongest driver of this transformation. These videos consistently ranked among the top-performing posts across platforms in terms of watch time, reactions and shares.
Digital PR and organisational engagement further amplified these results: collaborations with creators and institutional partners generated over 1.27 million additional views, extending reach beyond traditional audiences.

Conclusion
The evidence suggests that human-centred, researcher-led video formats—supported by AI-enabled production workflows—offer a concrete and scalable response to the challenges of the Age of Epistemia. By increasing transparency, recognisability and narrative clarity, these formats strengthen public trust, enhance comprehension and improve the dissemination of official statistics.