BelgiumPublic sector2011 – ongoingNational Social Security Office
Machine-learning risk models for social-fraud investigations
Belgium's social security office reported detecting complex fraud 2 quarters earlier with AI
Problem
Complex social fraud was becoming harder to detect through conventional field inspections. Cross-border schemes required investigators to identify relationships across large administrative datasets.
Approach
The office began AI experiments in 2011 and made predictive analysis part of its inspection work. It used supervised machine learning to assess risks from electronic records. Network analysis helped investigators identify links to employers previously associated with fraud.
Results
Takeaway
The models directed investigators toward suspected schemes. The reported earlier detection concerned complex fraud, rather than every inspection.
Categories
- Industry
- Public sector
- Business function
- Finance, risk and legal
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
The office described supervised learning and network analysis used in fraud investigations. It reported detecting complex fraud an average of 2 quarters earlier than before.
Full details are in the original sources.
