Powered by OpenAIRE graph
Found an issue? Give us feedback
ZENODOarrow_drop_down
ZENODO
Preprint . 2026
License: CC BY
Data sources: Datacite
ZENODO
Preprint . 2026
License: CC BY
Data sources: Datacite
versions View all 2 versions
addClaim

Who Pays the Cost of Verification? The Economy of Judgment in AI-Mediated Scientific Production

Authors: Viliotti, Andrea;

Who Pays the Cost of Verification? The Economy of Judgment in AI-Mediated Scientific Production

Abstract

La produzione scientifica mediata dall'IA viene oggi letta soprattutto in due modi: come accelerazione produttiva o come rischio epistemico. Questo articolo sostiene che nessuna delle due cornici basti da sola, perché entrambe sotto-specificano l'ecologia istituzionale entro cui le claim scientifiche vengono prodotte, verificate, contestate e archiviate. Basandosi su due lavori precedenti — uno che ricostruisce i diritti cognitivi come diritti di habitat e uno che introduce l'economia del giudizio come meccanismo a tre layer operante attraverso sei assi diagnostici — il saggio applica quel framework a un singolo caso documentato: il ramo pubblico GDE di gravità emergente e i suoi due companion manuscripts. La tesi è intenzionalmente stretta. Il paper non decide sulla correttezza della fisica sottostante e non offre misurazione causale né metriche validate. Chiede piuttosto se il framework aiuti a spiegare una redistribuzione strutturale in cui i costi di produzione diminuiscono mentre i costi di verifica vengono spostati a valle sulle comunità disciplinari, sulle infrastrutture archivistiche e sulle altre istituzioni del giudizio scientifico. L'analisi mostra che i sei assi — tempo di interpretazione, visibilità, provenienza, capability, contestabilità e pluralismo istituzionale — si applicano alla catena di produzione scientifica senza modifiche ad hoc. Il saggio introduce inoltre il concetto di distributed capability deficit, cioè una condizione in cui nessun partecipante della catena di produzione può certificare autonomamente l'output, e lo collega al rischio di sophisticated wrongness. Quattro proposizioni di medio raggio, tre hard nulls a livello di framework e una agenda di portabilità rendono l'argomento falsificabile. Il contributo è diagnostico, non prescrittivo: offre un modo analiticamente portabile per studiare la produzione scientifica assistita da IA in domini diversi, mantenendo espliciti i limiti di auto-riferimento, opacità di provenienza e validazione esterna.

AI-mediated scientific production is increasingly discussed either as a productivity breakthrough or as an epistemic risk. This paper argues that neither frame is sufficient on its own because both under-specify the institutional ecology through which scientific claims are produced, verified, contested, and archived. Building on prior work that reconstructs cognitive rights as habitat rights and introduces the economy of judgment as a three-layer mechanism operating through six diagnostic axes, the article applies that framework to a single documented case: the public GDE emergent-gravity branch and its two companion manuscripts. The claim is intentionally narrow. The paper does not adjudicate the underlying physics and does not offer causal measurement or validated metrics. It asks whether the framework helps explain a structural redistribution in which production costs decrease while verification costs are shifted downstream onto disciplinary communities, archival infrastructures, and other institutions of scientific judgment. The analysis shows that the six axes—time of interpretation, visibility, provenance, capability, contestability, and institutional pluralism—map onto the scientific production chain without ad hoc modification. It also introduces the concept of distributed capability deficit, a condition in which no participant in the production chain can independently certify the output, and links it to the risk of sophisticated wrongness. Four middle-range propositions, three framework-level hard nulls, and a portability agenda are formulated to make the argument falsifiable. The contribution is diagnostic rather than prescriptive: it offers an analytically portable way to study AI-assisted scientific production across domains while keeping self-reference, provenance opacity, and external-validation limits explicit.

Keywords

economy of judgment AI-assisted science verification costs interpretive authority large language models epistemic infrastructure cognitive habitat scientific production sophisticated wrongness provenance distributed capability deficit peer review open access authorship norms AI governance science and technology studies

  • BIP!
    Impact byBIP!
    selected citations
    These citations are derived from selected sources.
    This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    0
    popularity
    This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
    Average
    influence
    This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    Average
    impulse
    This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
    Average
Powered by OpenAIRE graph
Found an issue? Give us feedback
selected citations
These citations are derived from selected sources.
This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Citations provided by BIP!
popularity
This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
BIP!Popularity provided by BIP!
influence
This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Influence provided by BIP!
impulse
This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
BIP!Impulse provided by BIP!
0
Average
Average
Average
Upload OA version
Are you the author of this publication? Upload your Open Access version to Zenodo!
It’s fast and easy, just two clicks!