Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

High-Risk AI in Life and Health Insurance Underwriting - Balancing Underwriting with Data Protection and Non-Discrimination

Golding, Benjamin LU (2025) LAGF03 20252
Department of Law
Faculty of Law
Abstract (Swedish)
Sedan framväxten av artificiell intelligens (AI) har de teknologiska framstegen snabbt ökat i både intensitet och omfattning, ofta i en takt som överstiger den hastighet med vilken regelverk kan upprätthålla de rättigheter på vilka Europeiska unionen (EU) är grundad. Försäkringsbranschen är en sektor där teknologiska framsteg inom AI är avgörande för lönsamhet. EU har försökt begränsa potentiella kränkningar av mänskliga rättigheter vid användningen av AI-system genom att införa AI-förordningen. Regelverket har delvis varit föremål för negativ kritik, särskilt från försäkringsbranschen.

Uppsatsens fokus snävar in på definitionen av ”AI”, dess rättsliga beståndsdelar samt tekniska funktioner på vilka de rättsliga förutsättningarna vilar.... (More)
Sedan framväxten av artificiell intelligens (AI) har de teknologiska framstegen snabbt ökat i både intensitet och omfattning, ofta i en takt som överstiger den hastighet med vilken regelverk kan upprätthålla de rättigheter på vilka Europeiska unionen (EU) är grundad. Försäkringsbranschen är en sektor där teknologiska framsteg inom AI är avgörande för lönsamhet. EU har försökt begränsa potentiella kränkningar av mänskliga rättigheter vid användningen av AI-system genom att införa AI-förordningen. Regelverket har delvis varit föremål för negativ kritik, särskilt från försäkringsbranschen.

Uppsatsens fokus snävar in på definitionen av ”AI”, dess rättsliga beståndsdelar samt tekniska funktioner på vilka de rättsliga förutsättningarna vilar. När denna grund väl är lagd och kombineras med en undersökning av hur AI-system används inom försäkringssektorn, analyserar uppsatsen därefter bestämmelserna i AI-förordningen om högrisk-klassade AI-system. Den undersöker också hur försäkringsbolag, särskilt vid riskbedömning och prissättning av liv- och hälsoförsäkringar, lämpligen bör anpassa sin användning av sådana system utifrån dessa rättsliga ramar för att uppfylla rätten till dataskydd och rätten till icke-diskriminering.

Uppsatsens slutsatser omfattar de praktiska svårigheterna att inventera det dataunderlag och de steg – både tekniska och övergripande – baserat på vilka ett AI-system grundar sina slutsatser. I kombination med systemets frekventa benägenhet till bias och diskriminering innebär dessa faktorer regleringsmässiga och rättighetsbaserade utmaningar. Med teknisk kompetens, förståelse för regelverken och anpassning till de tekniska förutsättningarna i de aktuella systemen, tillsammans med en vilja till efterlevnad och innovation, kan dessa utmaningar emellertid hanteras. (Less)
Abstract
Since the inception of artificial intelligence (AI), technological advancements have rapidly increased in intensity and scope, often outpacing the speed at which regulatory frameworks can preserve the rights on which the European Union (EU) is founded. The insurance industry is a sector in which technological advancements are vital for profitability, a fact seemingly exponentially compounded with the rapid development of AI. The EU has tried to curb the potential infringements on human rights in the deployment of AI systems by introducing the AI Act. This regulatory act has been subject to criticism, especially from the insurance industry.

This thesis narrows in on the definition of AI, its legal constituent parts, and the practical... (More)
Since the inception of artificial intelligence (AI), technological advancements have rapidly increased in intensity and scope, often outpacing the speed at which regulatory frameworks can preserve the rights on which the European Union (EU) is founded. The insurance industry is a sector in which technological advancements are vital for profitability, a fact seemingly exponentially compounded with the rapid development of AI. The EU has tried to curb the potential infringements on human rights in the deployment of AI systems by introducing the AI Act. This regulatory act has been subject to criticism, especially from the insurance industry.

This thesis narrows in on the definition of AI, its legal constituent parts, and the practical functionalities on which the legal prerequisites are based. When that foundation is laid and combined with an examination of how AI systems are employed in the insurance sector, the study proceeds to analyse the ordinances in the AI Act regarding high-risk AI systems. It also examines how insurance companies, particularly in the case of risk assessment and pricing for life and health insurance, should preferably adapt their use of such systems based on these legal frameworks to comply with the right to data privacy and the right to non-discrimination.

Findings include practical difficulties of tracing the data input and steps, both technical and overarching, by which an AI system operates and reaches its conclusions. Combined with the system’s frequent tendency toward bias and discrimination, these factors provide regulatory and rights-based challenges. However, with regulatory understanding, adaptation to the technicalities of the systems in question, and a willingness for compliance and innovation, these challenges can be addressed. (Less)
Please use this url to cite or link to this publication:
author
Golding, Benjamin LU
supervisor
organization
course
LAGF03 20252
year
type
M2 - Bachelor Degree
subject
keywords
EU law, private law, AI Act, GDPR, Automated decision making, data privacy, non-discrimination, AI, AI-systems, profiling, credit-scoring, sensitive data, contestability, explainability, transparency
language
English
id
9216989
date added to LUP
2026-02-09 15:03:37
date last changed
2026-02-09 15:03:37
@misc{9216989,
  abstract     = {{Since the inception of artificial intelligence (AI), technological advancements have rapidly increased in intensity and scope, often outpacing the speed at which regulatory frameworks can preserve the rights on which the European Union (EU) is founded. The insurance industry is a sector in which technological advancements are vital for profitability, a fact seemingly exponentially compounded with the rapid development of AI. The EU has tried to curb the potential infringements on human rights in the deployment of AI systems by introducing the AI Act. This regulatory act has been subject to criticism, especially from the insurance industry.

This thesis narrows in on the definition of AI, its legal constituent parts, and the practical functionalities on which the legal prerequisites are based. When that foundation is laid and combined with an examination of how AI systems are employed in the insurance sector, the study proceeds to analyse the ordinances in the AI Act regarding high-risk AI systems. It also examines how insurance companies, particularly in the case of risk assessment and pricing for life and health insurance, should preferably adapt their use of such systems based on these legal frameworks to comply with the right to data privacy and the right to non-discrimination.

Findings include practical difficulties of tracing the data input and steps, both technical and overarching, by which an AI system operates and reaches its conclusions. Combined with the system’s frequent tendency toward bias and discrimination, these factors provide regulatory and rights-based challenges. However, with regulatory understanding, adaptation to the technicalities of the systems in question, and a willingness for compliance and innovation, these challenges can be addressed.}},
  author       = {{Golding, Benjamin}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{High-Risk AI in Life and Health Insurance Underwriting - Balancing Underwriting with Data Protection and Non-Discrimination}},
  year         = {{2025}},
}