Can You Trust the Machine? Artificial Intelligence Challenges the Insurance Industry to Re-evaluate the Role of the Expert

José Manuel Dias da Fonseca, Chairman and CEO of Brokerslink and CEO of the Portuguese MDS Group

Artificial intelligence is being utilized in an increasing number of stages across the insurance industry’s value chain, from risk assessment and underwriting to claims handling and customer advisory. At the same time, the question of when an AI-generated response can be trusted—and when human expertise and judgment are still required—is becoming all the more critical.

This was one of the central topics at the Brokerslink Global Conference 2026, which gathered 480 insurance and risk management professionals from 96 countries in Istanbul, Turkey, from November 4 to 6. The three-day conference program featured 15 events and 40 speakers. Themes included artificial intelligence, climate change, health, aviation, and sports-related risks.

Firstbrokers was represented at the conference by International Account Manager Eduard Kanoshin. In his view, one of the most interesting presentations at the event was Markus Leippold’s talk on the opportunities and risks of AI. Leippold serves as a Professor of Finance at the University of Zurich, a Senior Chair at the Swiss Finance Institute (SFI), and a researcher at Google DeepMind. He leads the SUREAL Lab (Sustainable and Resilient Economy AI Lab) research group at the University of Zurich and also works as a researcher at the University of Oxford. His research focuses on applying machine learning to climate risks, sustainable finance, and asset pricing.

A Convincing AI Answer Is Not Necessarily Correct

The title of Markus Leippold’s presentation, Trust the machine? AI and the risks that cannot be priced, encapsulated a fundamental question for the insurance industry. AI is already being offered as a solution for underwriting, claims handling, risk assessment, and customer advisory. However, the key issue is no longer what AI is capable of, but rather how the reliability of AI-generated information can be evaluated.

Leippold approached trust as a practical question. A particular challenge arises in situations where an AI-generated answer is difficult to verify independently. A convincing-sounding answer can look identical regardless of whether it is right or wrong. Therefore, when utilizing AI, it is essential to identify what can be left to the machine, what needs to be verified, and where an expert’s own judgment remains indispensable.

This question is particularly topical in the insurance sector. Decisions are often based on vast amounts of data, yet they simultaneously involve uncertainty, exceptional circumstances, and risks whose full consequences cannot be predicted or priced. AI can help analyze data and identify patterns, but it remains the expert’s responsibility to assess how much trust can be placed in that analysis in a given situation.

Leippold researches the application of machine learning to climate risks, sustainable finance, and asset pricing, among other areas. At Google DeepMind, his work focuses on leveraging AI in climate and scientific assessment. His current research examines human-AI collaboration specifically in questions where the right answer cannot be easily verified independently.

International Network Brings Knowledge to Firstbrokers’ Finnish Clients

The Brokerslink Global Conference is the flagship annual event of the international Brokerslink network. In addition to knowledge sharing, the conference aims to strengthen collaboration within the network and help local insurance brokers serve internationally operating clients. In Istanbul, brokers met with risk management professionals, insurance companies, experts, and researchers from around the world. Brokerslink’s international network currently operates in 142 countries and connects approximately 50,000 industry professionals.

The network’s international cooperation is also visible in practical development work. In connection with the Istanbul conference, Brokerslink launched a new digital platform, ekko, designed to make it easier to find specialized expertise among brokers in different countries and to manage cross-border client accounts.

The AI theme highlighted from the conference by Eduard Kanoshin reflects a broader shift in the insurance industry. While the importance of AI is growing, it is matched by an increasing emphasis on the expert’s ability to critically evaluate information, understand a client’s individual risks, and recognize situations where a machine’s answer alone is not enough.

It is precisely in this combination—the opportunities offered by technology and human expertise—that the true value of artificial intelligence in the insurance industry will largely be determined.

—

Firstbrokers Oy was founded in 1989 and is thus the first insurance brokerage company in Finland. Customers are Finnish SMEs and subsidiaries of international companies. Firstbrokers Oy is also the most international insurance broker in Finland, as it is a member of several global insurance broker networks. This enables business customers to provide risk management, insurance and employee benefits services easily, cost-effectively and reliably in more than 110 countries. Firstbrokers Oy covers its customers’ business and employees in all industries. Firstbrokers Oy continuously manages the risk management, insurance and employee benefits of more than 200 companies, and has already served more than a thousand companies during its 35 years of operation. We’ve got you covered!