Artificial intelligence (AI) is one of the key technologies in the digitalization of the insurance industry. However, consumers see the development of artificial intelligence as cause for concern.
Systems with artificial intelligence increase efficiency and therefore profitability in insurance companies. Techniques such as machine learning enable automated processing of routine queries and claims settlement. And they support marketing and product management in the development of new products and contribute to a better understanding of customers.
German consumers, however, are worried about AI according to a survey conducted last year.
The customer data paradox
The survey conducted last year on behalf of the "Verbraucherzentrale Bundesverband" (VZBV)paints a similarly paradoxical picture to that seen in other sectors.
On the one hand, many people share personal information with strangers, for example, on social networks such as Facebook, without any hesitation, even though it is now widely known that this data is used for profiling and advertising. And in the worst case, this also gives criminals enough information to be more successful in phishing attempts.
On the other hand, there are concerns about the (automated) analysis of data. In surveys, customers regularly ask retailers for more personalization options and a more personal approach. But as Forrester Research found in 2019, an imaginary line exists here between what customers find useful and measures they consider "creepy".
This attitude, which is sometimes hardly rationally comprehensible, is also adopted by the respondents towards AI systems in the insurance industry.
AI is viewed critically
Insurance companies are by no means allowed to implement everything that AI systems make technically possible. This is because the current data protection law, the strict regulations within the insurance industry and ultimately the constantly expanding jurisprudence set narrow limits here.
Apparently, however, this is not enough for consumers. 75 percent of those surveyed by the vzbv are concerned about automated decisions within companies.
The study participants did not condemn these processes in general, but view the systems critically when it is unclear to them what data and principles are being used. Without this information, algorithms are seen as a "black box" and their use is a cause for concern.
In view of these fears, it is not surprising that only 18 percent of those surveyed see more opportunities than risks in the use of automated systems when algorithms make decisions or influence decision-making processes.
"The diffuse fear of a machine taking on a life of its own is unfounded in the European insurance industry," clarifies Karsten Schmitt, Head of Business Development at adesso insurance solutions. "AI used in claims processing, for example, automates tasks that were previously performed by employees. Of course, the same laws for the protection of personal data apply as for manual processing. No insurer will take the risk of breaching the obligation to protect personal data. The consequences in terms of penalties and loss of reputation are too great for that."
Transparency and benefits against concerns
The maximum demand of the VZBV and the respondents is that the state should be given the possibility to check such algorithms in order to ensure compliance with the law.
But aside from the ongoing debates about ethics in AI or government oversight of development, simple measures could make for a better image of automated AI.
Since the people concerned seem to lack confidence in the technology, it seems appropriate to approach the topic of AI on two levels:
On the one hand, the benefits of AI need to be communicated more clearly. The use of a chatbot offers more flexibility for policyholders. For example, you can get the information you need faster. Automated AI solutions can speed up the regulatory process, claims are compensated more quickly, which should be in the interest of those affected. Fraud detection systems save money for all policyholders.
At the second level, more transparency should be created about the use of technologies. To take the example of chatbots: It must be clearly communicated when there is contact with a machine system instead of a human being. In informational brochures and notices, insurance companies should not rely solely on data protection notices and general terms and conditions or contractual conditions, but should communicate directly which data is used automatically for decision-making. Explanations in simple language could prove to be an effective lever.
It is in the nature of things that transparency finds its limits where it threatens a competitive advantage. The importance of AI, however, is too great to hide, because it can, for example, reduce complexity and make service more customer-oriented. Advantages from which the insured parties benefit and which can reduce mistrust and fear. Karl-Heinz Streibich, President of the German Academy of Science and Engineering, appropriately writes in a guest article for Wirtschaftswoche, a leading German business magazine "That is why it is important to make AI understandable and to promote dialogue with the public. Ultimately, it's safe and ethical AI systems that people trust."
Find out why insurance companies are choosing to use AI and how it is done successfully in our free whitepaper (German only).