Do we need new rules for ar­ti­fi­cial in­tel­li­gence in the in­sur­ance in­dus­try?

Interview

The SIA takes a position and recommends six AI principles for insurance. Sandra Kurmann, Divisional Head of General Operating Conditions at the SIA, summarises the sector recommendations for the use of AI in a brief interview and explains what consumers can expect from the sector in the future. 

Artificial Intelligence (AI) is already part of everyday life for consumers, whether as helpful chatbots on insurance company websites or in the background during the claims settlement process. Consumers often do not know exactly where AI is hiding or what kind they will encounter in the future.

 

Sandra Kurmann

It must be ensured that the use of AI is declared to the customer: Sandra Kurmann, Divisional Head of General Operating Conditions at the SIA.

How important is AI in the insurance sector? How is it being used today?

AI is used in various forms in the insurance sector. The areas in which it is applied can be grouped into four categories: first, to optimise/automate repetitive internal processes, particularly in claims settlement; second, to improve risk assessment, which means better forecasting, using, for example, actuarial models; third, to enhance the customer service experience, using chatbots such as ChatGPT; and fourth, to develop new products thanks to new data sources and suitable distribution channels, such as insurance for flight delays.

It is important to note that the strengths of AI currently lie primarily in highly repetitive tasks and where analysis of large data volumes is required. However, the actual decision is still made or validated by a person. 

 

How does the insurance sector uphold its digital responsibility? Is AI really essential?

The tight regulatory framework within which insurers operate places the protection of policyholders at the centre. Insurers must make sure that the use of AI is declared to the customer. The usage of data in particular is subject to comprehensive rules that ensure data protection, transparency and traceability, and the avoidance of discrimination or manipulation. Data protection legislation sets an appropriate framework for handling data and protecting privacy. 

Artificial Intelligence has the potential to make the processes within the insurance value chain more efficient and accurate. So, businesses will need to use it to be successful and to stay competitive. If they do not make use of such advanced technology, insurance companies will lose their market share in the medium and long term. 

«People will continue to take priority as decision-makers in the insurance industry.»

What are the risks of AI for insurance customers? Why are recommendations needed for the use of AI in the sector, and what does the SIA expect to gain from this?

Customers sometimes worry that they might lose all their privacy as policyholders due to the relatively large amount of data available and the use of AI. A reality check is necessary here. When it comes to AI, it is always a matter of weighing up what is technically feasible, legally permissible and socially acceptable. Insurers are aware of their role and responsibility towards customers and adopt numerous measures in the area of corporate governance in order to supplement the existing tight regulatory framework and adequately address any ethical issues. With this sector recommendation, the private insurance sector wishes to make this transparent to the outside world, not only to the regulator and supervisory authority, but also to customers. An insurer's business model is the assumption and handling of risks – and this careful handling of risks is also evident in the use of new technology. 

 

Insurance is a complex matter for many consumers. That is why there is a fear that in the very near future no one will be left in insurance, only robots. Is this a foreseeable development?

It is the job of an insurance company to assume risks and in so doing to strengthen the financial resilience of an individual and of the entire economy. The focus is on the customer and their needs. Even if appropriate AI tools can assist employees and take on some of the workload, people will continue to have primacy as decision makers in the insurance sector and are thus indispensable. This is particularly true in claims settlement. Customers who want personal advice will certainly always be able to find someone at an insurer who will give them personal and individual advice.

How does the insurance sector compare with other industries in terms of digital responsibility? 

The Ethos Study 2022 identifies seven principles of digital responsibility, including establishing digital governance, ensuring transparency and complying with data protection and ethical principles. The industry comparison reveals that the insurance sector performs best in these principles, just as it did in the previous year. Private insurers are already taking numerous steps to address the points mentioned, particularly in the area of corporate governance. The SIA’s sector recommendation also contributes to this point.

SIA recommendation on the use of Artificial Intelligence (AI)

Date: September 2023

  1. Primacy of human actions
    The insurance company focuses on people as decision makers and ensures appropriate control of systems and algorithms using suitable mechanisms. The results achieved by AI can also be checked and traced by the organisation.
     
  2. Social and environmental well being
    When using AI, the insurance company ensures that data is handled in a measured and sustainable manner that creates added value for customers and stakeholders.
     
  3. Protection of privacy, data quality management and transparency
    The insurance company ensures compliance with data protection legislation, adheres to an appropriate standard of quality in respect of the origin and processing of data, and ensures that individual automated decisions meet transparency requirements.
     
  4. Fairness
    The insurance company implements appropriate mechanisms to ensure fairness in the application of AI procedures. In doing so, it complies with actuarial requirements (e.g. statistical tests).
     
  5. Governance and responsibility
    The insurance company creates governance mechanisms appropriate to the use case that ensure that AI algorithms and systems are assessed in terms of their risk, responsibilities are defined and quality assured.
     
  6. Technical robustness and security 
    The insurance company regularly evaluates situations in which AI algorithms may not or no longer function properly and addresses such situations promptly and transparently, thereby ensuring the technical robustness and security of the systems.