The two institutions conducting this survey are the Alan Turing Institute, the UK’s leading centre for research expertise on AI and data, and the Ada Lovelace Institute, experts in public engagement around such issues. The survey is to be funded by two of the UK Government’s research councils. All this points to the survey being given a lot of attention when it comes out.
So why should insurers be particularly interested? These two paragraphs from the announcement illustrate why…
Participants will be asked about a range of specific uses of AI, from facial recognition and medical diagnostics to driverless cars and credit scoring. This will provide richer and more detailed insights than simply asking about ‘AI’ as an abstract concept.
The survey will generate a rich dataset providing policymakers, researchers and practitioners with robust evidence to help them ensure that AI and data-driven technologies work for people and society.
Just from this synopsis, it’s clear that insurers can expect to see outputs from the survey that will be relevant to life and health underwriting (re medical diagnostics) and personal lines underwriting, claims and counter fraud (re credit scores and emotional AI like facial recognition).
Oven Ready for Legislators
The fact that the survey organisers are after “…richer and more detailed insights…” point to the findings being made more ‘oven ready’ for policy makers to form legislative proposals and to question regulators about where they stand on the issues and what they’re doing about them. Credit scores in particular are very likely to be a concern (more here).
This is highlighted by reference in the announcement to the “…potential benefits and risks, as well as attitudes to AI governance and regulation.” Given worrying signs of regulatory drift on data and analytics in financial services (more here), the survey findings could come as something of a jolt to the FCA’s own attitudes.
Insurers need therefore to prepare for a knock on reaction from the regulator, channelled principally through expectations of the scope and depth with which firms are implementing the consumer duty. In essence, the political stakes relating to the FCA needing to make the consumer duty a regulatory and political success are on a continual rise (more here in respect of discriminatory pricing).
More than Just Pot Half Full
Another area of insight that the survey will provide is around digital poverty, as the third of these bullets points from the announcement highlights. Bodies like the Treasury Committee often take an interest in issues like this and I would expect what the survey finds will be explored in depth by them. Again, this will matter for insurers’ product design, distribution and governance.
- a focus on specific technology use cases
- a survey design that will explore why people currently think the way they do about AI
- a representative sample of people who are currently offline (without internet access)
- a specific focus on AI rather than data
I’ll end with a brief comment about the second of those four bullet points. We know from the research commissioned a few years ago by the Association of British Insurers (more here) that consumers view insurers’ handling of their data through what the authors called ‘a double lens of mistrust’. This new ATI / ALI survey may well provide insight that will allow insurers to differentiate between ‘why people think the way they do about AI’ and ‘why they think the way they do about AI and insurance’.
Have insurers conducted research like this themselves? It would be great to hear from any that have.