Il saggio intende mettere in luce gli impatti che i processi di digitalizzazione, big data e intelligenza artificiale hanno avuto nel settore assicurativo sia nella produzione dal punto di vista del disegno del prodotto, sia nella distribuzione attraverso sistemi di profilazione dei clienti, sia nella vigilanza con possibili ispezioni mirate.
The paper considers the impact of digitalization processes, big data, and artificial intelligence in the insurance market with regard to production and products design, customers profiling in distribution and better addressed supervision.
Keywords: insurance – new technologies
The term Insurtech refers to the application of digital technologies to the insurance world. The areas of application range from production to distribution to insurance governance itself. From the production point of view, digital technologies have affected the insurance world due to the new coverage needs mainly due to data security. On the production side, the real news concerns the application of blockchain technology to insurance contracts. An innovative frontier on the production side could be represented by open insurance. The insurance industry 4.0 includes all the technologies that have led to a digitization of relationships, accelerating their establishment and execution, facilitating automation processes also thanks to the use of algorithms . In the insurance sector, digitalization has allowed the exploitation of data collected from customers together with big data to perform clustering operations capable of profiling customers and improving the adherence of products to their insurance needs. The use of big data becomes important in the insurance industry. As it is well known, the term “big data” indicates an enormously large complex of data that can be used to form new knowledge through the relationships between knowable data. This information, due to its size and speed of acquisition, has a heuristic value as it represents the starting point for identifying correlations that may be relevant for future developments . There are different techniques used: “Data mining” is the process of analyzing data from different points of view to obtain useful information. It is the process of searching for correlations or patterns between data collected in relational databases. “Data fusion” is the process of integrating multiple data and knowledge. The expectation is that the “merged” data contain information that is superior to the original data. The “clustering” procedure aims to group data and organize them into groups so that the data contained in the same cluster are more similar to each other than those contained in different “clusters”. The “regression analysis” is used to estimate the strength and direction of the relationship between the variables that are in linear relationship to each other. Eiopa (European Authority on Insurance and Pensions Funds) in its Report “Report on Best Practises on Licencing Requirements, Peer-to-Peer [continua ..]
From the production point of view, digital technologies have affected the insurance world due to the new coverage needs mainly due to data security. On the production side, the real news concerns the application of blockchain technology to insurance contracts. An innovative frontier on the production side could be represented by open insurance. Awareness of cyber risk is growing and with it the search for tools to deal with it. For their part, insurers seem to have identified a sector of activity, but they have also taken note of some critical issues that make it difficult to apply traditional risk management models to the so-called cyber risk. For both policyholders and insurers, technology can be a disruptive element or a driver for development. Everything will depend on the resilience capabilities of the market, operators, and individuals. Cyber risk is well on the international agenda. The accessibility, reliability and security of cyberspace were considered by the G7 Leaders’ Summit in 2016 as an “essential foundation for the economy, growth and prosperity”. Insurance companies and insurance intermediaries, in the management of policyholders’ data, will be one of the subjects on which civil or administrative liability may be imposed pursuant to the GDPR. One of the great opportunities for insurers, however, is represented by the possibility of placing on the market functional products to cover cyber risk. Insurance contracts can represent an answer, not only in terms of insurance coverage but also in terms of risk management tools and implementation of prevention systems. The cyber risk, due to the a-spatial and a-temporal characteristics within which it develops, will hardly be able to find solutions when it is realized in the production of a concrete damage that can also activate an interminable chain of losses and events. Damage compensation is not the answer. Cyber risk is solved by measures to prevent and contain harmful effects. Let’s start by saying that the mapping of risks and the identification of insurance coverage depends on the qualification of the insured party also in this sector. In the case of private citizens, the risks are generally: On-line reputation damage. Digital identity theft related to credit. Online / e-commerce purchases. Cyber crime, in particular the theft of money and valuables through computer fraud. With reference to companies, [continua ..]
Insurance distribution sees a progression of procedural steps marked by documentation, information, registration, and communication obligations ordered to the suitability of insurance contracts respect to the request and needs of insured party. The fulfillment of these obligations could be guided and recorded, also for the purpose of conservation and proof of correct execution, through the use of block chains. Block Chain could also facilitate to find the most suitable products on the market. Art. 20 of IDD (directive 97/2016 on insurance distribution) requires that “any proposed contract must be consistent with the insurance requests and needs of the customer” . As above mentioned, Blockchain works as a decentralized and encrypted register, in which, in real time, countless operations are recorded without anyone being able to change what is written centrally, but any modification or update can only take place after receiving the consent. by all parties involved in the transaction to be registered or modified. The Blockchain therefore could be considered as a third player, potentially replacing the functions that today we usually attribute to notaries who have been identifying ways of using it for the purpose of exercising the notary profession for some time. Blockchain allows to collect, verify and share data of various kinds in a safe and transparent way. These data may include the demands and needs of customers, the results of their profiling and product information . As for customer profiling and product adequacy determined through algorithmic processes, it is necessary to distinguish: Ö in the life business, adequacy is more measurable. In the financial market, metrics are already in place that allow to quantify the adherence of the product to the profiles of adequacy and appropriateness with respect to the customer profile. Ö in the non-life branch, there are still no metrics, adequacy still finds a non-quantitative determination. It is a question of thinking about possible measurements in this area as well.
Insurance governance, as reviewed by the Solvency II reform (Dir. 138/2009), is based on a logic of analysis of the risks to which the assets of the insurance company are exposed according to a principle of ultimate responsibility of the directors who will have to decide based on data assumed by the various company functions. The requirements of prudent management , as in Solvency II, indicate the need for a management that can guarantee the solvency of the company through compliance with insurance techniques, efficient management of assets and adequate internal control and management procedure in order to give policyholders greater security and at the same time to guarantee a more effective management of the company, also in the perspective to obtain a higher degree of competitiveness between companies. The board of directors could use of Blockchain to collect data to take decisions. Some authors wonder if an artificial intelligence system could be part of the board. Commercial law scholars have recently dealt with the topic, showing the possible advantages and possible risks of conflict of interest of the operators who developed the algorithms that drive the artificial intelligence system. They propose solutions on a possible legal personality recognized to the A.I. that could be part of the board, in line with the theories that recognize the possibility of including legal persons among the members of the board of directors, or assuming to admit an IT services company to the board of directors .
Of course the technology can help the supervisors. In the last years Eiopa stimulated a discussion on the use of new technologies in supervision to deliver innovative and efficient supervisory solutions that will support a more effective, flexible, and responsive supervisory system. Eiopa identified the following strategic areas that can benefit from the use of technology in supervision activity and particularly: > improvement of data sharing and data analytics, in particular the development of a common risk assessment framework. > market monitoring from a conduct of business perspective, in particular by developing a tool to automate the assessment of the information available in the Key Information Document (KID) established by the Packaged Retail Investment and Insurance Products (PRIIPs) Regulationor in the Insurance Product Information Document (IPID) established by the Directive 97/2016. > assessment of existing Solvency II data in combination with available and comparable data on consumer complaints data with the use of new technologies, including the development of a Retail Risk Indicators Dashboard. In this regard it should also be considered to gather additional information on retail products relevant for conduct of business purposes.
Eiopa in January 2021 published a report on open insurance, a phenomenon of the share economy that can be seen from different perspectives also from the perspective of supervision . There is no uniform definition. EIOPA considers open insurance as accessing and sharing insurance-related personal and non-personal data usually via Application Programming Interfaces (APIs). Application programming interface (API) is a computing interface that defines interactions between multiple software instances or layers, including those operated by third parties. On the consumer data side, it could be defined as accessing and sharing data relating to consumer insurance services (e.g. insurance policy data such as insured item, coverage, claim history, Internet of Things data, etc.) between insurers, intermediaries or third parties for the creation of applications and services. On the supervisors’ side, open insurance could also open the doors to new supervisory tools. EIOPA has published a supervisory technology strategy (SupTech) explaining the use of technology by supervisory authorities to provide innovative and efficient supervisory solutions that will support a responsive supervisory system On the insurance industry side, the increased exchange of data through Open Insurance can facilitate industry-wide innovation, openness and collaboration and will likely enable the insurance industry to embrace data-driven innovation, the creation of innovative products for consumers and increase efficiency and interaction with third parties (e.g., better interaction with insurance platforms and ecosystems). Furthermore, it could facilitate the emergence of increased competition within the value chain such as new players, new business models, perhaps pushing down some costs through efficiencies. In the perspective of the industry angle should be considered also the interaction between banks and insurers considering the role of Bank-Insurance phenomenon Another angle to be considered should be the angle of international organizations interested in obtaining data, coming from the insurance market, for social purposes, like in case of health data. There are other possible implementations of open insurance not considered in Eiopa’s paper: 1. Index based indemnification Open insurance can have an important role in nonlife insurance in the procedures of damage assessment. The determination of losses can take long periods of time and could be costly for both [continua ..]
Insurtech can represent a turning point in the insurance industry by shifting the insurance business more on the front of risk management than on that of its coverage: Ö On the non-life front, the insurer can help the insured to reduce risks through digitalization processes that allow the implementation of an iron curtain of preventive mechanisms Ö On the life front, the insurer can help customers improve the allocation of their financial resources by optimizing results. The use of technology can find some legal constraints. For instance, in Italy the data sharing finds an obstacle in the activity of the Market Authority as it is possible to consider the agreement on data sharing an anticompetitive agreement. This point needs a clarification at European level. In our opinion the data sharing cannot obstacles the competition considering the difference between data and actuarial models. Only models can be considered the know how of an insurance company. Of course, a common regulation of those phenomena is needed to avoid lack of uniformity and competition in Eu. The increased data sharing requires revisions in the regulatory framework related to insurance data, especially on the side of licenses The best licensing approach is creative common license with a strong regulation. As said A CC license is used when an author wants to give other people the right to share, use, and build upon a work that the author has created. An interesting CC license is the Attribution Share Alike CC BY-SA image. This license lets others remix, tweak, and build upon your work even for commercial purposes, as long as they credit you and license their new creations under the identical terms. This license is often compared to “copyleft” free and open source software licenses. All new works based on yours will carry the same license, so any derivatives will also allow commercial use. This is the license used by Wikipedia, and is recommended for materials that would benefit from incorporating content from Wikipedia and similarly licensed projects. It is important to consider also ethical issues relevant in light of open insurance. In addition to what EIOPA said in its paper , we see particularly problem of discrimination. Big Data Analysis by insurance firms could lead to hyper personalised risk assessments, leaving certain and possibly broader segments of consumers 'uninsurable'. It is important to find form of coverages for those people. In [continua ..]