Conceptual essay reveals the lack of judgment provenance in automated data systems, highlighting the need for attributable human oversight in AI decisions.
Hans Rosling's Factfulness (2018) established data as the antidote to instinct. By 2021, critics noted that data alone was insufficient — yet offered no alternative. By 2026, AI-generated "doctors" on YouTube had reached, by one independent count, more than a billion views across dozens of languages, targeting elderly audiences with no accountable authorship. This essay identifies a common blind spot across these moments: the absence of provenance of judgment. Statistics aggregate; provenance attributes. Knowing what happened is not the same as knowing which human decided, what they assessed, and what they took responsibility for. As a conceptual extension of the factfulness project, the paper proposes preserving attributable, examinable records of human judgment alongside system outputs in AI-assisted decisions — an approach related to Human Oversight Evidence (HOE). This article is a conceptual essay, not an empirical demonstration that provenance improves decision quality, truthfulness, or regulatory compliance.
No takes yet. Share an insight, caveat, or question.
Michael Bugay (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: