Research project AI to detect unclear insurance claims
Can AI help us come to terms with fraud? In this project, we use machine learning models to identify indicators for unclear insurance claims in order to deter suspicious damage payments and insurance fraud.
The process of identifying unclear insurance cases is complex and require manual oversight. The identification and selection of unclear insurance cases are dependent on the claims officer’s diligence.
In the absence of technical tools for handling large amounts of data, identification and selection is made largely based on the claims officer’s intuition. The selection of unclear insurance cases is likely to be affected by, for example, detected historical frauds at the expense of things that characterize ongoing fraud. When the claims adjuster has identified an unclear insurance case, an investigator takes over.
The purpose of this research project is to develop machine learning models that can, with high precision, identify indicators of unclear insurance cases.
Photo: Andrey Popov/Mostphotos.
6. People in the project
- Contact person at SU
Isak Samsten (https://www.su.se/profiles/iska9819-1.191333)
– felaktiga uppgifter BTW ska vara Senior lecturer och inte postdoc
- Other project manager/s
Joakim Lundin, Länsförsäkringar AB
- Other member/s of the project
Steven Gruenhut, Länsförsäkringar AB
8. Funding
Länsförsäkringsbolagens Forskningsfond
https://www.lansforsakringar.se/stockholm/privat/om-oss/hallbarhet--forskning/forskning/om-forskingsfonden/
9. External collaboration/s
Länsförsäkringar AB
Project members
Project managers
Isak Samsten
Senior lecturer
Joakim Lundin
Members
Steven Gruenhut