No. 54

Share

Print

Well-Being in Estonia 
Is Generally Good — 
But Not for Everyone

20 November 2026

Focus

RiTo No. 54, 2026

Algorithmic tools can help welfare institutions allocate limited resources, but their predictions may hide contested assumptions about who deserves support.

Using the example of employment-service decision support and a participatory exercise, the authors ask whose understanding of fairness is embedded in data and models. Predicting someone’s employment prospects is not the same as identifying their needs. Fairness requires transparency, participation and accountable human judgement.

Feedback