ARTIC data mining lifts undeclared-work reassessments ninefold

The Cour des comptes found that reassessments from checks suggested by data mining rose ninefold at URSSAF Île-de-France. They went from €3.7 million in 2023 to €33.5 million in 2024.

  • 2 an organization

    Reassessments from checks suggested by data mining at URSSAF Île-de-France rose from €3.7 million in 2023 to €33.5 million in 2024.

Depth runs from 1, a convenience or time saved, to 5, a life or a livelihood preserved.

Status
Reported and intended benefits.
Evidence

foundstudiedreportedclaimed

An independent audit report supports the benefit.

Compared with

Reassessments from checks suggested by data mining at URSSAF Île-de-France in 2023. Measured.

Without the system
3,7 M€ in 2023
With it
33,5 M€ in 2024
The system’s part
Not isolated. The Cour compares amounts from checks that data mining suggested in one region over two years. ARTIC reached the network only in 2024, and the Cour neither separates ARTIC from other data mining nor compares these checks with others. In a footnote, the Cour writes that the data were rebalanced so that the algorithm would learn to tell firms apart.
Limits
Not yet checked.
Costs and harms
Not yet checked.
Later
Not yet checked.

Download data Cite Suggest a correction

Benefit type
Improved Task Accuracy
System
ARTIC (Analyse des risques pour le travail d'investigation et le ciblage), the URSSAF data-mining tool
Made by
Agence centrale des organismes de sécurité sociale (ACOSS)
Used by
URSSAF network (figures from URSSAF Île-de-France)
Where
Île-de-France, France
Stage
In routine use
Date
Kind
Aggregate result
Broad benefit
Analyzing Data or Information, Evaluating Information to Determine Compliance with Standards
Collections
Scrutiny and findings
Last revised
31 October 2026

Sources

  1. Audit report, by authors with no tie to the system.

    Cour des comptes. La politique de lutte contre la fraude aux prélèvements sociaux liée au travail dissimulé. Évaluation de politique publique, communication au comité d’évaluation et de contrôle des politiques publiques de l’Assemblée nationale, April 2026, pp. 76-77.

    Les montants de redressements issus de contrôles suggérés par le datamining ont été multipliés par neuf entre 2023 et 2024.

Details

AI system

Workflow

ARTIC scored firms' risk from criteria such as declaration frequency, wage inconsistencies and past disputes. It drew on monthly payroll declarations and on data from partners such as Tracfin and the tax services.

Follow-up

Conditions for replication

Inspectors must take up the tool's suggestions when they choose checks. The training data must include compliant firms so that the algorithm learns to separate them from likely fraudulent ones.

Entry details

Sector

Social contributions enforcement (undeclared work).

Purpose

The ACOSS built ARTIC to score firms' risk of undeclared work and map their links to other entities, both for choosing checks and during investigations.

Not yet checked: model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.

Cite this record

Peter H. Vartanian. (2026). ARTIC data mining lifts undeclared-work reassessments ninefold (AIBD-0688). Artificial Intelligence Benefits Database, edition of 31 October 2026. https://petervartanian.xyz/aibd/records/AIBD-0688.html (archived at https://doi.org/10.7910/DVN/AADSII, version 4.0)

BibTeX Data Field definitions