Personalisation is the big structural trend happening in insurance at the moment. And it has attracted all sorts of arguments around it in relation to fairness. Some people emphasise the fairness of personalisation (why pay for other people’s risk / claims / etc), while others stress the unfair outcomes that can be associated with it (access, bias, autonomy, etc).
The introduction of digital fairness legislation in the EU is likely, as a side consequence, to draw out the key arguments around how profiling and personalisation are being applied in insurance. The sector would be fielding their best arguments in support of what digital strategies are delivering, while consumer groups would be challenging this and evidencing the impacts being experienced. It’s happening already (for example, EU versions of the UK’s poverty premium), but would move to a new level with a Digital Fairness Act.
Just a EU thing though? Well, yes and no. It will only affect insurers doing business in the EU, but at the same time, just like the GDPR and the AI Act, it will send out waves of influence to other countries and regions.
The question I want to consider here relates to whether or not a Digital Fairness Act (DFA) would significantly help to resolve the fairness issues associated with profiling and personalisation in insurance. First however, a quick look at what the DFA is looking like.
What are the Concerns?
The intention is for the DFA to strengthen European online consumer protection rules by tackling:
“…unethical techniques and commercial practices related to dark patterns, marketing by social media influencers, the addictive design of digital products and online profiling, especially when consumer vulnerabilities are exploited for commercial purposes.”
It’s judged to be necessary because of gaps in existing legislation and fragmentation of relevant national laws in various EU member states. How it will exactly take shape though is still uncertain. It may become an ambitious single piece of legislation, or end up as a modest package of amendments to existing law.
A lot of the debate over the next year or so about how the legislation is to be shapes will be around the interplay of growth and competitiveness, and consumer protection. There are many advocates for the former, but the latter is still very strongly supported.
The latter can draw upon some pretty grim findings from recent surveys. For example, a 2023 survey by the EU found that 70% of consumers were concerned about how their personal data was used and shared. And the EU’s Fitness Check survey in 2024 found that 37% of consumers thought that a company had knowledge of their vulnerabilities and used it for commercial purposes.
Is It Actually Needed?
Let’s address this question specifically in terms of insurance, and in relation to the profiling and pricing practices associated with personalisation.
I think ‘is it needed’ is not perhaps the right question to start with, in the context of insurance and profiling. That’s because so much of the DFA will be addressing non-insurance things. What this means then is that how those non-insurance aspects of profiling and pricing are addressed will almost certainly sweep over and affect insurance itself.
For example, this could be in relation to how the DFA defines and scopes fairness, what it means by profiling, and so on. These core aspects of a piece of fairness legislation will have strong knock-on effects into insurance.
Will those knock-on effects help or hinder the debate around fairness and personalisation now gathering pace in insurance? I think it could do both. It will help draw together and shape consumer concerns, and give them a powerful voice. It will also shake up the sector a bit and perhaps make them realise that some form of ‘new thinking’ is needed.
On the other hand, it could hinder the fairness and personalisation situation in insurance at the moment. That’s because I believe that legislation on fairness per se could move that ‘insurance situation’ forward a bit, but then cause it to be stuck short of what insurance needs.
Value Driven
So why is that likely to be the case? Two reasons. Firstly, it is unlikely that fairness will ever be truly ‘nailed down’ and agreed in sufficient detail by all the parties engaging with the drafting of a piece of legislation. And secondly, fairness is a more complicated thing when it comes to insurance, than it is in terms of retailers or manufacturers. Just think of actuarial fairness and moral hazard for starters.
So why can’t fairness be nailed down in sufficient detail? That’s because it is too value driven, too socially and politically dynamic, for consensus to be reached through a legislative drafting process. We have to accept this. What is needed instead is an agreed form of engagement for people to negotiate and reach that consensus on what is fair in this or that situation.
I explored this in some depth in a 2023 paper for the Institute and Faculty of Actuaries, called ‘Revolutionising Fairness to Enable Digital Insurance’ (read it here). After much research and numerous discussions, the feedback I got from academics was to look at the history and use of common pool resources (CPRs). CPRs have been used for many years to reach a consensus around the fair use of natural resources.
Could CPRs be used in relation to fairness and a financial resource like insurance? Let’s say I didn’t think that it would be impossible, just a bit challenging. A lot of people told me that it would be impossible. I understand their foreboding, but then, I’ve always thought that if business people and communities could come together around a common pool resource to manage water extraction in southern California in the 1950s, then who else couldn’t do likewise! If nothing else, the processes by which CPRs emerge and take shape will be very informative for whatever mechanism of engagement and agreement is adopted for fairness and insurance.
Unintended Consequences
Legislation is an expression of power by representatives of the people, based around the social values in ascendency at the time. What goes into legislation is mostly determined by whoever’s values carry most sway during the legislative drafting process. It is still though a political process, and the output often reflects that. It is how society has organised itself, and by and large, it has proved to be successful.
That said, legislation often has unintended consequences, and with the Digital Fairness Act, one unintended consequence could be to firstly, stir up the ‘fairness and personalisation’ situation in insurance, and secondly, land it short of a working consensus, and leave it stuck there. In short, forwards, to a degree, but no solution.
The future of insurance, interlinked as it is with issues like fairness, will not be determined by this or that stakeholder, even one with the power of legislators. It will be determined by the engagement of multiple stakeholders, working together to address the key social justice issues that digital insurance is raising. The best role for legislators is in fact to stand back a bit and ensure that those that need to be at the table are there, that the agenda is right and the table is level. It would be quite a new thing if the Digital Fairness Act were to deliver this.
Let’s end with an analogy. A crucible can mean a vessel for melting metals, but another meaning is as "a place or situation in which different cultures or styles can mix together to produce something new and exciting.” The future of insurance needs less of a digital fairness act and more of a crucible, to produce an digital insurance that is new, exciting and fair.
