Data and Power pt4 – the right skills are vital to success in data ethics

As data and analytics transforms insurance, the sector’s demand for people with digital skills has increased enormously. Yet if insurers are to really get to grips with data ethics and ensure it is part of how ‘we do digital round here’, they must take a close look at the type of skills they’re prioritising. A commitment to data ethics can only succeed if the people delivering it are skilled not just in data science, but the social sciences as well. In this fourth and final post in the ‘data and power’ series, I’ll be examining the crucial role played by education.

For every article about how exciting the digital transformation of insurance is, there’s seems to always be a parallel article raising concerns about the sector recruiting enough of the right people to deliver it. And both types of article are right: the future of insurance is exciting, and getting the people to deliver it is a big challenge.

In response, there has arisen a veritable production line of degrees and courses to meet the growing demand for data scientists. And within insurance itself, we’re seeing a lot of reskilling going on. Zurich for example recently announced that it is to retrain two thirds of its UK workforce over the next five years, in order to ‘future proof’ their workforce.

The Right Skills

And this is of course all good stuff, so long as the scope and depth of skills being learnt is right. All too often, digital courses focus on the individual mastery of abstract and technical concepts. These concepts are important, but so equally are the ethical concepts that underpin an insurer’s commitment to data ethics, and that support individual and corporate accountabilities for regulatory obligations like fairness, honesty and non-discrimination. As pricing people in insurance have found, those obligations have hard boundaries, even when data offers up all sorts of opportunities and insights.

I would go so far as to say that delivering on a commitment to data ethics will be nigh on impossible unless the people involved understand data and analytics in the more rounded way I refer to above. Remember that survey from 2018, in which 28% of insurance executives thought the biggest risk they faced was from personal regulatory sanction because of their firm’s compliance failings. I believe that nearly all significant compliance failures over the next five years will be linked to how firms are using data and analytics. The Financial Ombudsman Bureau thinks so too.

Four Skills Examples

So what are the skills that need to be incorporated into this reskilling? This post is not the place to list them all, but what I would say, is that people building this digital transformation of insurance need to be equipped to recognise and manage the big questions around ethics, values and power that their work is clearly now raising. Let’s see this through four examples drawn from the themes we’ve covered in recent ‘data posts’.

• Digital relationships – the ability to recognise and build relationships, so as to engage with and learn from the key audiences that influence, and are influenced by, your digital strategies.
• Discrimination and access – the ability to recognise why and how data and analytics are not neutral and objective, so as to address this in processes like data cleansing, categorisation and significance levels.
• Data and due diligence – the ability to recognise and account for the different contexts within which data arises, so as to integrate this into due diligence, projects and audit processes.
• Visualisation and reporting – the ability to configure output data in relation to the people who experience the outcomes.

The obvious question that then arises is around who should be learning such skills? The digital and transformation teams are obvious candidates, but another way to see it is through the question of whose work is being influenced by this digital transformation. The answer that produces is, well, just about everyone in the firm, from recruitment to planning, from claims people to board people.

Early Stages

How should we judge where insurers are now in relation to steps like the four set out above? We can look at the insurers who have appointed customer experience directors, and strived to put together diverse digital teams. And those are good steps, but they’re still only a beginning. As we saw in this earlier post, there is still a need for diversity in thinking, in perspective, in experience.

I’m talking here about digital teams having people able and empowered to think differently and to challenge, for example, what a dataset might be missing, and why. It needs people who are as skilled in the social sciences as in the data sciences, in order to see things that digital technologists often do not fully grasp. And it needs people who have experienced what it means to live on the digital margins, to draw attention to missed stories.

When you consider these types of skills, I believe that there are few, if any, insurers who do not still have a long way to go. And while I have not been able to locate any data or survey material about the composition of digital teams, I can instead look at some of the outcomes that data science has been so influential in producing for insurance. Those outcomes around pricing, those behind concerns about discrimination, those around the use of certain types of data, are indicative of digital teams in the insurance sector still in need of step change improvements.

Getting Started

What might a committed insurer do now? I would suggest it starts by looking in three areas: the learning strategy and provision, performance design, and recruitment processes. And who should do that ‘looking’? I would suggest three people: an internal person, an external digital skills person and a critical friend drawn from a consumer group. The last of those three might seem like the odd one out, but based upon what we’re covered in this ‘data and power’ series, they could well be the most important.