Discriminatory welfare risk-scoring algorithm
In October 2024, La Quadrature du Net (LQDN) along with 14 other organisations began litigation against a risk-scoring algorithm used by the CNAF (Caisse nationale d’allocations familiales), the family branch of France’s social security system. It manages social benefits of more than 12 million people, including housing and minimum-income benefits.
The case builds on pre-litigation research which showed that the algorithm is highly discriminatory. LQDN assesses that socio-economic variables have a preponderant weight in the calculation of the score, structurally disadvantaging people in precarious situations. The algorithm functions like a “double penalty”: it targets those who are most oppressed or marginalised already.
Discriminatory welfare risk-scoring algorithm
Organisation Name
La Quadrature du Net
Country/Jurisdiction
France
Grant Amount
Litigation Support: EUR 30,880
Pre-litigation Support: EUR 20,303
Current Status
Pre-litigation research complete; litigation ongoing
Image credit: Bastian Riccardi on Pexels
Grant type
Litigation Track Support
Grant type
Pre-litigation Research Support
Description
In October 2024, La Quadrature du Net (LQDN) along with 14 other organisations began litigation against a risk-scoring algorithm used by the CNAF (Caisse nationale d’allocations familiales), the family branch of France’s social security system. It manages social benefits of more than 12 million people, including housing and minimum-income benefits.
The case builds on pre-litigation research which showed that the algorithm is highly discriminatory. LQDN assesses that socio-economic variables have a preponderant weight in the calculation of the score, structurally disadvantaging people in precarious situations. The algorithm functions like a “double penalty”: it targets those who are most oppressed or marginalised already.
"The algorithm functions like a “double penalty”: it targets those who are most oppressed or marginalised already"
Strategic Goal
To shed light on the development and harms of such algorithms that are now spreading to other administrations under the justification of public services digitalisation.