Occasional Paper Technology Towards 2033

Technology Towards 2033 The Future of Insurance and Supervision Read more Occasional paper

2 Contents Executive Summary 3 01 Introduction 5 1.1 Objective 5 1.2 Scope and Approach 5 1.3 Guide 6 02 Technological Trends 7 2.1 Data 7 2.2 Data Storage/Access 8 2.3 Data Analysis 9 2.4 Connectivity 10 2.5 Digital Security 10 2.6 Digital Skills & Legacy 11 2.7 Conclusion 11 03 Impact on Insurance 12 3.1 Process Optimization 12 3.2 New Propositions 14 3.3 Changing Roles 17 3.4 New Dependencies 18 3.5 Other Developments in Distribution and Product 18 3.6 Conclusion 19 04 Impact of EU Policy 20 4.1 Horizontal Policy Developments 20 4.2 Digital and Data Strategy of the European Commission 22 4.3 Sectoral Policy Developments 22 4.4 Conduct Supervision in the EU & Cross-border Issues 22 4.5 Conclusion 23 05 Impact on the AFM 24 5.1 Role of the AFM in the Whole 24 5.2 Public Interest, Risks, and Supervisory Approach 24 5.3 Conclusion 28

Executive Summary 3 Digitalization will undoubtedly have an impact on the insurance sector and the supervision thereof. Insurers are acquiring more data, applying smarter data analyses, and thus influencing their entire business operations and distribution, from pricing to claims handling. This applies not only to Dutch insurers, but to many European parties that may also become active in the Netherlands. This brings new questions and challenges for supervision. This report looks at the possible influence of a digitalizing world with a ten-year horizon. We outline the most important technological trends, the impact of these on insurers and distribution, the influence of European legislation, and the impact of this whole on the AFM and policymakers. The most important technological trends range from an increased amount of data and greater analytical power, to increasing connectivity and issues regarding security and skills. The number of data sources is growing, inter alia due to internet-connected devices, making continuous (data) interaction possible between the insurer and the customer. Numbers and text are enriched with analyses of conversations, photo and video material as the main data sources. Connectivity and the ability to open up and share data allow for new collaborations and services that were previously not possible. Insurers can apply these techniques in their pricing, terms and conditions, and underwriting policy. Every consumer can receive a customized premium, which either reflects their refined actuarial risk or their willingness to pay. Policy conditions no longer need to consist of fixed sets, but can be tailored to measure. In this way, every consumer receives a fully individualized product with a corresponding price. The underwriting policy can be refined as much as insurers can set it up. Difficult-to-insure risks, such as cyber threats, can remain affordable through prevention. The handling of claims, including fraud detection, can be automated to a large extent. The distribution of insurance is changing under the influence of embedded insurance and robo-advice. By integrating insurance modules into websites, for example ('embedded'), they do not only have to be offered when purchasing a product, such as a trip or bicycle, but distribution can extend to a potentially infinite number of intermediaries, during the lifetime of a product. But also insurance policies that are separate from a product, such as life insurance, can be concluded via various platforms. Robo-advice makes it possible to advise on complex and impactful products without human intervention. In embedded insurance, but also via the classic distribution channels. Data usage facilitates the development of new types of insurance. It opens access to insurance that is time- and location-bound, such as during work on a flex contract or a short visit abroad. Behavioral pricing can be used for almost all products, ranging from rewarding driving behavior and lifestyle, to reading out devices connected to the internet. Parametric insurance, where the insurer pays out if a certain threshold value is reached (such as the predetermined amount of rainfall at a festival), is easier to develop when the insurer has more data, but may also be necessary to keep certain risks insurable. In addition to the influence on processes, distribution, and products, the role of the insurer itself could change significantly. In the platform role, an insurer becomes increasingly dependent on additional services. This can relate to a product, such as breakdown assistance with a car insurance, but can also fall outside this, such as installing solar panels. A more advanced scenario is the insurer as the director of an ecosystem, which can take care of a customer from A to Z on themes such as vitality or housing. Executive Summary

4 The role of the AFM will change along with developments and risks. For example, risks regarding solidarity and uninsurability mean that the AFM must have an even stronger signaling role towards policymakers, in addition to looking at individual institutions. Integrating insurance into a product, service, app, or website potentially leads to a multitude of intermediaries, which may require a new supervisory approach. Supervision of suitable advice changes when the by-products for consumers are more decisive. The importance of cooperation with other supervisors increases: within the Netherlands, and within Europe. However, the current starting points of the AFM remain unchanged in a certain sense: forward-looking supervision on both processes (such as PARP) and outcomes (such as advice). In this way, we safeguard individual and collective financial well-being, and a level playing field. We will have to align our supervision as much as possible with future developments in the coming years. In addition, we call on insurers, involved supervisors, and policymakers to think about the possible implications of the outlined expectations, inter alia in terms of ethics and policy. In order to reduce the risks of digitalization in the future, and at the same time to seize the opportunities to improve products and services for consumers. An alternative is that the insurer gets a smaller role in the entire chain, whereby the control and contact lie with the manufacturer or an established platform party (such as a BigTech). Reinsurers can also indirectly compete with the insurer, by supporting insurtechs financially or with insurance expertise. In both variants, it is essential that the level playing field does not become unbalanced due to a data advantage obtained from market power or the concentration of one or more players relative to the insurer. Europe is trying to channel technological developments in the right direction with both horizontal and sectoral legislation. The series of regulatory initiatives from Europe emphasize the need to counter challenges arising from the digitalizing world and to protect the position of the consumer. In addition to the new obligations that these initiatives bring with them, they also open up new possibilities (for example through data sharing), which can also influence the insurance sector. The expected developments under the influence of digitalization come with new risks. For example, personalized pricing and conditions can lead to a breakdown of solidarity and uninsurability. Insurers operating as a platform or ecosystem increasingly offer services and products that are not subject to supervision in principle, but can be decisive for customers. The increasing possibilities for cross-border service provision, facilitated by digitalization, emphasize for the AFM the importance of conduct supervision that is high on the agenda in all European countries and is interpreted uniformly. Executive Summary

5 1.1 Objective With this research, we try to answer four questions, looking at the world of 2033: • What does digitalization mean for the insurance sector and more specifically for the market parties that the AFM supervises? • What is and will be the influence of regulation from Europe? • What does this mean for the way the AFM supervises? • What ambitions must be formulated now to be prepared for the possible reality in ten years? This exploration aims to provide insights and points of attention for supervisors, insurers, and policymakers. 1.2 Scope and Approach This exploration focuses on the full breadth of insurance products and distribution. With regard to stakeholders, the focus is on the parties that currently fall under the supervision of the AFM. These are insurers, authorized representatives, and intermediaries/advisors. Knowing that the ecosystem and the relationships within that ecosystem can change, other stakeholders will not be lost sight of. The potentially changing role and relationship between the parties in the ecosystem will therefore be addressed in this exploration. The insights from this exploration stem from a combination of literature research and conversations with stakeholders within and outside the sector. The concept of insurance is centuries old and, according to stories, even goes back to 500 BC. The need to be able to cover risks has only grown over the centuries. For consumers, the types of insurance can roughly be divided into two themes: damage and life. Many of the insurances within these themes have remained largely unchanged for years. The question is to what extent and how the digitalizing world will have an impact on insurance. The above suggests that the concept on which insurance is based (covering risks) will not be affected quickly. There will always be risks and new risks will arise, for which coverage will be sought. It is possible that certain risks are reduced or completely solved by digitalization, making insurance of those risks unnecessary. At the same time, digitalization will undoubtedly have an impact on the insurance sector. This can be thought of as embracing digital technologies to optimize processes and develop new propositions, shifts in the value chain, and possible changes in the role of the currently supervised institutions. Ultimately, it is possible that the core of the insurance concept changes, for example in terms of solidarity. The question is also whether this requires a change in supervision and to what extent the role of the AFM relates to other supervisors in the digital playing field. Within the concept of insurance, existing risks can be exaggerated and new risks can arise. The AFM will continue to strive for compliance with the duty of care and the centralization of customer interest. 01 Introduction 01 Introduction

01 Introduction 6 1.3 Guide With this research, we have tried to sketch a provocative image of the possible future of insurance in 2033. First, an overview is given of the technological developments that may have an effect on the sector (Chapter 2). Next, the possible impact of these developments on the sector is described (Chapter 3). (The prospect of) the impact of technological developments and the associated changes lead to a diverse palette of EU policy (Chapter 4). All this has consequences for the role of the AFM in the large supervisory landscape and for the way supervision is exercised (Chapter 5). Figure 1. Technological trends and the impact on insurance, policy, and the AFM Impact on Insurance Impact of Policy Impact on the AFM Tech trends insurance 2033 Supervisory approach Demonstrability Role of AFM in whole Point of attack Horizontal policy Sectoral policy Conduct supervision Other propositions Changing roles New dependencies Other developments Process optimization Data/IoT Data storage/access Data analysis(power) Connectivity Internal External DLT APIs Cloud AI Satellite 5G/6G Cybersecurity

7 Data Sources Where organizations were limited to the data they generated themselves or obtained from interaction with (potential) customers at the advent of the internet, the data web has expanded at a high speed. The data used to guide consumers in the customer experience process, from orientation to purchase, is accessed from various sources. For each phase of the customer process, the optimal data sources are identified and combined to be used via (almost) real-time analysis. Initially, attention was focused on attracting and onboarding customers, but insurers can also focus more strongly on the monitoring process in the coming years. The own data building of a customer starts with onboarding. Where necessary, external data sources are used, as this promotes data quality and/or the ease of onboarding. The data flow that arises in the monitoring process depends on the type of insurance, the business process, and the interaction between insurance and customer. Here lies a strong added value of data and this offers opportunities to make adjustments that are valuable for the customer and insurer. Smart, internet-connected devices will take off in the coming years and contribute to the continuous interaction and data flow between customer and insurer. This data flow offers opportunities to better understand customers and translate the insights into further personalization of, for example, products, prices, and services. What developments form the insurance market in the next ten years? What is the specific impact of technology and digitalization? In this chapter, we describe developments that influence the insurance market. We focus especially on the impact of technology and digitalization. Data is increasingly generated by devices related to health, fitness, mobility, and the home environment. Cars sometimes resemble mobile phones, there is no watch that only tells the time anymore, and in the house, devices can be controlled remotely via an app. Wifi is no longer just for the laptop, and 5G and satellite networks are also used by cars. This offers convenience, but also has challenges. The economy and society grow along with the technological developments outlined. This can lead to complicated situations, for example in the field of privacy. Who exactly receives the data that my car, watch, or coffee maker produces? 2.1 Data The digital system is built to generate, store, share, process, and present data. With every action on and from a device, data is generated and exchanged. Data can be obtained in multiple ways. It can be so-called own data that has been developed internally, resulted from analyses, or obtained during an interaction with a (potential) customer. Data can also be obtained externally through collaborations, openly available sources, purchases, or devices. Here, the Internet of Things (IoT) plays an important role, where everyday devices such as a coffee maker or baby monitor are connected to the internet. 02 Technological Trends 02 Technological Trends

02 Technological Trends 8 An increasingly important question is where the data lies. As described, data is generated in more and more places. Gaining access to the data that is relevant for an institution is becoming increasingly essential. Institutions will therefore have to ask themselves what data is available, what data is relevant, and how they can gain access to that data (such as purchase, collaboration)? A challenge for the future will be to have and maintain an overview of the data available at that time. The data curve continues to grow exponentially and the quality can improve. But data can also become outdated, lose relevance, or prove to be incorrect. Data Forms The times of numbers and plain text as the main data forms are over. Advanced analysis techniques make it possible to record and analyze other forms of data. Conversations can now be translated into text, making automated analysis of conversations possible. Neural networks can analyze photo and video material at the pixel level, offering a range of possibilities in a sector where visual evidence, especially in the damage segment, is very valuable. Data Standardization Data standardization is essential to make data shareable with a larger group of (third) parties. Standardization of the data and the access, for example via Application Programming Interfaces (APIs), can only be enforced by legislation. Data Sovereignty Data was called the new gold for years. In exchange for user data, 'free' services and apps were offered on the internet and via the smartphone. The relationships can shift in the coming years (among other things under the influence of legislation), whereby the consumer gains more control and insight into the use of his or her data by organizations. Organizations will have to respect the boundaries of data use and become more transparent about the use of data for services or processes. Data Quality and Reliability Data quality has a domino effect on all subsequent steps in the process. Without the highest data quality and reliability, models lose their quality and added value, and incorrect outcomes can have disastrous consequences. Organizations will be keen to ensure that data quality can be guaranteed and that as much as possible can be drawn from raw source data. At the time of writing, there are no quality marks that guarantee data quality. Data quality and reliability are closely linked. The rise and popularity of smart devices and sensors also contain a strong dependency. A sensor can, for example, register certain activities or specific behavior, but a defect will lower the reliability of the data. How can the consumer defend himself if the box in his car records speeding violations that were never committed? 2.2 Data Storage/Access Cloud In a digital system, all produced and collected data must be stored somewhere. Initially, this was possible on internal storage media. The exponential growth with which data is produced, including improved connectivity, means that cloud solutions are now used to store data.

02 Technological Trends 9 The cloud seems to become the central and beating heart of the digitalizing organization. It is the place where data, analysis, computing power, and access possibilities come together under a strongly protected roof. It is the cloud where data is received, processed, and accessed to other sources. At the same time, standard and custom analyses can be performed, and extra computing power can be added quickly. The business model of cloud providers is not focused on cloud space, but on the services offered around it. Think of standard analyses, extra computing power, and possibly cloud computing services in the near future. In a digital environment where it is desired or required by regulation to share data with third parties, cloud providers can provide data management environments in which organizations maintain overview and control over the data flows in and out of the organization. DLT/Shared Infrastructure1 As data becomes commonplace, and the driving force of digital solutions, there will be a need to share it more broadly. Private and public parties will seek each other out in so-called data ecosystems, supported by a shared infrastructure, to easily share data that contributes to a process or business model. APIs APIs are software interfaces that make it possible for different applications to communicate with each other. This makes the transfer of data from one system to another possible. The technology is not new and was previously mostly used to let internal systems of organizations communicate with each other. For a few years now, the technology has also been used to provide access to data to third parties, allowing new services to be created. APIs form the springboard to make data within one organization available to another organization, allowing new services and services to be developed. The concept works both ways. 1 IAIS Report on Fintech developments in the insurance sector (iaisweb.org) Thus, an organization can attract data via API solutions, but also make data available to external parties. 2.3 Data Analysis Artificial intelligence is a word that has existed since the second half of the 20th century. In recent years, this technological development, like the underlying techniques such as machine learning and deep learning, has taken off enormously. This is partly due to increased computing power, low costs of data storage, and ever-growing data streams to feed AI. In the coming years, insurance products, partly driven by AI, can develop from reactive (detect and repair) to proactive (predictive and preventive). AI can be applied in every process in the chain, with the aim of optimization, in whatever form. Themes such as explainability and ethics will remain relevant. The use of historical data and the manner of modeling continue to raise points of attention, such as the chance of extrapolating historical patterns, the awareness of biases, and the fact that models are always a simplification of reality. The ethical issues are relevant for every step in the process (input, throughput, output), and the need for explainability remains a direct and indirect consequence thereof.2 The AFM drew attention to this in 2019 in collaboration with De Nederlandsche Bank (DNB) in its exploration of the use of Artificial Intelligence in the insurance sector.3 Access to more and more data points of an individual and the increased possibilities of data analysis make it increasingly possible to segment at the micron level. Digital environments and the digitalization of behaviors in the physical world (for example sensors, video/photo) make it possible to analyze and follow behavior at the individual level. 2 Towards a Standard for Identifying and Managing Bias in Artificial Intelligence 3 Artificial Intelligence in the Insurance Sector – An Exploration

02 Technological Trends 10 Satellites The continuous increase in the number of smart devices and IoT solutions places high demands on connectivity possibilities