188 M Tanninen, T-K Lehtonen, and M Ruckenstein
how the notice-and-consent model, utilised, for instance, in the General
Data Protection Regulation (GDPR) enforced in the European Union, is an
inadequate means to ensure trustworthy data practices.
Experimenting with digital data requires insurers to leave what appeared
to be the ordered world of ‘pure’ and insulated statistical information in
which they are comfortable operating. Although insurance has never been
only about statistical data and actuarial calculations (Ericson & Doyle 2004;
McFall 2014; O’Malley & Roberts 2014; Van Hoyweghen 2007), the ability
to amass and use longitudinal data sets has been a self-evident characteristic of insurance companies to the degree that these operations have been
normalised. Data has been defined by certainty in the sense that its uses
and movements have been strictly regulated and predictable. However, with
the new operations, insurers face novel uncertainties that involve regulatory
instability and data existing ‘in the wild’ because it flows in the ‘real world.’
Before they can wholeheartedly embrace these new developments, insurers
need to experiment with the promise they offer. Even if the data cannot be
fitted into neat actuarial categories and statistical analyses, it is seen as a
potential new tool and resource, whose value lies in correlations, probabilities, and predictions. Furthermore, it is hoped that digitally tracing what
people do will give insurance companies visibility into their lives and offer
the possibility to gently manipulate or ‘nudge’ (Thaler & Sunstein 2009) customers’ everyday behaviour in a direction that would be more cost-efficient
for insurers in the long term.
As we will demonstrate, however, all this requires that the new practices
are seen as valuable and trustworthy by policyholders. If entering the messy
realm of digital data is a leap of faith for insurance companies, it is equally
so for their customers. Paradoxically, although insurance is intended to provide security and mitigate risk, it can create new anxieties and uncertainties
for the consumers (Booth & Harwood 2016). Insurance is an opaque technology to begin with, and the actual trade-offs of a given contract are difficult
to estimate. Behaviour-based insurance further complicates the insurer–
policyholder relationship, as activity data collected by smartwatches and
smartphones and lifestyle interventions aim to gently push people towards
healthier and safer habits. In other words, even if people’s daily lives are
already permeated by messy data practices in the realms of digital services,
retail, and social media, creating new kinds of relationships with an entity
like an insurance firm is far from straightforward.
To shine a light on how existing and potential policyholders see insurers’ attempts to form relationships with them through personalised data
collection, we analyse issues raised by data use through a case study of
two Finnish behaviour-based life insurance policies. Our main aim is to
discuss the uncertainties related to data practices. These uncertainties, we
argue, are fundamental to understanding the contextual nature of datafication processes. Obtaining value out of digitalisation requires that data
flows can be secured; people need to trust that the operations will benefit
how the notice-and-consent model, utilised, for instance, in the General
Data Protection Regulation (GDPR) enforced in the European Union, is an
inadequate means to ensure trustworthy data practices.
Experimenting with digital data requires insurers to leave what appeared
to be the ordered world of ‘pure’ and insulated statistical information in
which they are comfortable operating. Although insurance has never been
only about statistical data and actuarial calculations (Ericson & Doyle 2004;
McFall 2014; O’Malley & Roberts 2014; Van Hoyweghen 2007), the ability
to amass and use longitudinal data sets has been a self-evident characteristic of insurance companies to the degree that these operations have been
normalised. Data has been defined by certainty in the sense that its uses
and movements have been strictly regulated and predictable. However, with
the new operations, insurers face novel uncertainties that involve regulatory
instability and data existing ‘in the wild’ because it flows in the ‘real world.’
Before they can wholeheartedly embrace these new developments, insurers
need to experiment with the promise they offer. Even if the data cannot be
fitted into neat actuarial categories and statistical analyses, it is seen as a
potential new tool and resource, whose value lies in correlations, probabilities, and predictions. Furthermore, it is hoped that digitally tracing what
people do will give insurance companies visibility into their lives and offer
the possibility to gently manipulate or ‘nudge’ (Thaler & Sunstein 2009) customers’ everyday behaviour in a direction that would be more cost-efficient
for insurers in the long term.
As we will demonstrate, however, all this requires that the new practices
are seen as valuable and trustworthy by policyholders. If entering the messy
realm of digital data is a leap of faith for insurance companies, it is equally
so for their customers. Paradoxically, although insurance is intended to provide security and mitigate risk, it can create new anxieties and uncertainties
for the consumers (Booth & Harwood 2016). Insurance is an opaque technology to begin with, and the actual trade-offs of a given contract are difficult
to estimate. Behaviour-based insurance further complicates the insurer–
policyholder relationship, as activity data collected by smartwatches and
smartphones and lifestyle interventions aim to gently push people towards
healthier and safer habits. In other words, even if people’s daily lives are
already permeated by messy data practices in the realms of digital services,
retail, and social media, creating new kinds of relationships with an entity
like an insurance firm is far from straightforward.
To shine a light on how existing and potential policyholders see insurers’ attempts to form relationships with them through personalised data
collection, we analyse issues raised by data use through a case study of
two Finnish behaviour-based life insurance policies. Our main aim is to
discuss the uncertainties related to data practices. These uncertainties, we
argue, are fundamental to understanding the contextual nature of datafication processes. Obtaining value out of digitalisation requires that data
flows can be secured; people need to trust that the operations will benefit