The (Un)Limited Use of AI Segmentation in the Insurance Sector

Financial Law Institute Working Paper Series 2023-05

24 Pages Posted: 10 Mar 2023 Last revised: 11 Apr 2023

See all articles by Julie Goetghebuer

Julie Goetghebuer

Ghent University - Financial Law Institute

Date Written: January 30, 2023

Abstract

In theory, under the premise of freedom of will, insurers in Belgium are permitted to employ artificial intelligence (‘AI’) and big data analytics for segmentation purposes, enabling them to determine the ‘claims probability’ (i.e., the likelihood and severity of potential claims) for each prospective policyholder. This analytical approach empowers insurers to determine whether or not they should underwrite a risk, and if so, on what terms. A more extensive pool of policyholder data may increase the accuracy of the assessment of claim probability and surpasses the information asymmetry between the policyholder and insurer.

Although the implementation of AI and big data benefits insurers, pervasive segmentation through AI has some negative implications and could entail serious ramifications for policyholders if their risk is incorrectly calculated. The current insurance regulation does contain some existing restrictions that limit the freedom of insurers to use AI; nevertheless, these provisions fall short of protecting policyholders from inaccuracies in their risk assessments and thus from receiving incorrect premiums or conditions.

Keywords: Artificial Intelligence, Big Data, Insurance Sector, Segmentation, Claims Probability, Risk Assessment, Freedom of Will, Liability, Policyholder Protection

JEL Classification: K22, K29, K10, K13

Suggested Citation

Goetghebuer, Julie, The (Un)Limited Use of AI Segmentation in the Insurance Sector (January 30, 2023). Financial Law Institute Working Paper Series 2023-05, Available at SSRN: https://ssrn.com/abstract=4381268 or http://dx.doi.org/10.2139/ssrn.4381268

Julie Goetghebuer (Contact Author)

Ghent University - Financial Law Institute ( email )

Ghent, B-9000
Belgium

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
65
Abstract Views
325
Rank
617,745
PlumX Metrics