Dario Sansone; Anna Zhu (2021) Working paper / technical note

Machine Learning in the Welfare System

Austaxpolicy: The Tax and Transfer Policy Blog

Ref: SRC-005-AUS-001

https://www.austaxpolicy.com/machine-learning-in-the-welfare-system/

Accessed: 3/23/2026

Summary

Author blog post providing accessible summary of the research. Identifies novel predictors (annual income variability, residential relocation frequency, mutual obligation non-compliance). Advocates for a system to monitor and audit automated decision-making. References Robodebt scandal as cautionary example. Proposes differentiated case management: light-touch for low-risk vs intensive for high-risk individuals.

View Harvard reference

Sansone, D. and Zhu, A. (2021) 'Machine Learning in the Welfare System', Austaxpolicy: The Tax and Transfer Policy Blog, 24 June. Available at: https://www.austaxpolicy.com/machine-learning-in-the-welfare-system/ (Accessed: 23 March 2026).

Attached File

FILE

SRC-005-AUS-001_Machine_Learning_in_the_Welfare_System.html 30 KB

Download

Linked Cases (1)

AUS-001 AUS Design & Development Phase

Machine-Learning Early Warning System for Income Support Recipients (Australia — Research Prototype)

Social assistance Classical ML