The UK Government sees location data as an important component of its digital strategy for the country. It published its Geospatial Strategy in June 2020 and in so doing, acknowledged that it will only succeed if location data is used in ways that retained public confidence.
The problem is that, to quote the Director of the Geospatial Commission (a Government body)…
“the necessary evolution of our geospatial ecosystem raises new and significant ethical considerations”
One outcome from that strategy document was the initiation of a dialogue with members of the public. The results of this public dialogue were published in November 2021 and can be found here.
The dialogue process looks pretty robust to me. And those drawing the conclusions from it look to be pretty independent from Government itself. This means the results are worth paying attention to.
Insurers are a Key Audience
And insurers should be a key audience for the dialogue’s findings. Information about the location of a risk influence all sorts of insurance decisions, from underwriting and counter fraud through to claims and marketing. On the face of it, it’s general insurance that seems to be most relevant, but dig deeper and it’s clear that life and health are very relevant too. Take the definition of location data used in the dialogue:
“Location data is any data or information that can describe the position of an object or person in relation to the Earth or some other object or person.”
You can see then that location data covers both data about people and data about objects and places. And if you think about it, it would be pretty strange if it didn’t.
So what conclusions did the dialogue draw, and how might they impact insurers? What’s the relationship between a place you travel to, and a place you move around within. Is location data personal data, perhaps even sensitive personal data? What steps should insurers take to align themselves with the direction that the Government’s geospatial strategy is taking? These are what I’ll look at next.
Five Issues Relevant to Insurance
These are five issues coming out of the public dialogue that I believe are relevant to insurers.
In broad terms, the dialogue found that people were broadly supportive of location data, seeing benefits in it for themselves and for society. That said, they also had concerns, in particular in relation to privacy, autonomy and inequality. What was interesting was the way in which benefits were most often framed as being for the wider public, while risks were most often framed around individual concerns.
The dialogue also found that the mix of benefit and concerns were influenced by the level of trust that the public had in the type of organisation involved. This then influenced perceptions about how ethical that organisation’s use of location data would be.
There was a recognition that the for-profit use of location data could be beneficial, but concerns were voiced about data markets and where their data could end up when it’s shared with third parties.
The level of detail at which location data was collected also raised concerns. Participants questioned why very specific data collection (tying a particular person to a particular place) was needed. They felt more comfortable with aggregate data. In weighing up the tensions between specificity and privacy, participants saw only a few circumstances in which they were willing to forego privacy. One example of this was when emergency services needed to know the whereabouts of a specific person.
Consent also mattered. Participants wanted the choice to consent or not consent to the collection of their location data.
What Does these Five Issues Tell Insurers?
In terms of societal benefits and individual concerns, insurers wanting to emphasise the former need to also account for the latter. Individual concerns can’t be discounted because of societal benefits. Both need to be part of your digital narrative.
We know from research commissioned in 2020 by the Association of British Insurers (ABI) that the public sees insurance through a ‘double lens of mistrust’. In other words, public trust in the sector is pretty low. This would therefore point to insurers needing to be more careful than most other sectors in their accessing and use of location data.
And it was also clear in that ABI research that the public were concerned about the extent to which consents were respected, and the extent to which insurers bought from, and sold into, data markets.
Finally, the granularity to which location data is collected should also be weighed up. Insurers may be able to collect data down to a per property basis (perhaps now in these IoT times on a per room basis), but should they? What are the incremental benefits and insights that greater granularity gives, versus the extent to which this undermines trust in the sector at both the individual and public levels?
Four Key Conclusions
I would urge insurers to pay attention to this research, and to the policy ramifications it could have. It is after all a Government document that is inter-twined with one of its key policy objectives. And I would suggest that insurers orientate their engagement around the dialogue’s four key conclusions.
- Make sure that your use of location data is inclusive. This means…
- don’t use it in ways (or to an extent) that certain communities or people feel ‘left out’;
- don’t let your for-profit benefits override the benefits and concerns of location data to individuals or wider society;
- use non-identifiable location data where ever you can.
2. Be accountable for how you use location data, and have structures and processes in place to deliver that accountability. This means…
- don’t gauge your progress just against compliance. Data protection legislation is not as reassuring to the public as firms may think.
- take steps to evidence how reliable your accountability is.
3. Be more transparent. This means…
- being more open about what location data is collected, what it is used for and who has access to it;
- produce this information in a more accessible and more usable form.
4. Enable agency. This means…
- don’t leave consumers with a feeling of disempowerment in relation to their location data;
- build more accessible, simple and honest mechanisms for consumers to engage with and making decisions around data collection and use;
- don’t see the collection of location data as something ‘done to’ consumers. Let it be more like something ‘done with’ them.
- only collect location data that is identifiable with clear choice, consent, accountability and transparency, otherwise you won’t be trusted.
A Particular Insurance Example
In a recent article looking at the fairness of individual rating factors, I looked at how Professor Kiviat’s research produced some interesting findings in relation to telematics data. Here’s a quote from that article…
“50% of respondents thought that it was somewhat or very fair for a motor insurer to use data about how a person drives (whether they slam on the brakes, turn sharply, etc.). Only about 30% of respondents thought that it was somewhat or very fair for a car insurer to use data about where or when a person drives. Americans think it is fair for insurers to hold them accountable for some, but not other, behaviours."
This points to two things. Firstly, this US research backs up the UK public dialogue on location data. And secondly, there are levels of granularity in the collection of data to which insurers are happy to go, but not the public. What insurers may find is that their current market and technological empowerment in relation to data collection may not last for long. Government policy could well intervene. That’s why I urged insurers in this recent article to make sure their systems are adaptable.
Looking Forward
The Geospatial Commission wants to continue its public dialogue process to explore how the concerns raised by this first round can be addressed. In doing so, I would urge the Commission to consider these questions:
- When does location data become personal data? And when does it become sensitive personal data?
- Data doesn’t sit on its own – it relies on algorithms to draw insight from it and other data sources associated with it. How can you frame the ethical issues in relation to this wider picture?
- How do you treat location data in terms of place, when that place moves from being on a per-property basis to being on a per room basis?
