Matthew Lehman's article is absolutely correct about the need for trust in insurance when using extensive personal data.

As AI becomes increasingly deployed in insurance (it is still a long way behind other industries) a big issues will emerge: Insurers deploying Transparent AI (where you can explain how actions were decided, with an audit trail) versus Opaque AI (where it is not explainable) will create a big issue with insurance regulators and customers alike.

Similarly GDPR, being enacted in May 2018 will provide an opportunity for insurers to prove themselves as trustworthy brands. It is a one-off opportunity for them to get clients to clean their own data and burnish their reputations. They should take it.