Every now and then IE likes to round up various insurance related news items from around the web. It’s like Readers’ Digest 2.0. Let’s chew on those tasty news nuggets.
BBC FRAUD
That would actually be a great Freeview Channel, but moving on, the Beeb is reporting on the IFB story regarding fronting by young drivers, and their exasperated parents.
It’s something that’s been going on for decades of course. But recently, premiums of over £9000 have been reported by some, with especially high quotes on pure EV vehicles too which are seen as intrinsically `good’ cars, despite the battery packs requiring exploitative labour practices in developing world mining and China based manufacturng plants.
Soaring premiums are potentially the biggest PR headache for the insurance sector this year and as the GE approaches, we can expect Martin Lewis and others to make the most of what they perceive to be `a rip off’ cartel of insurers milking the Motor market.
IE’s advice is arm yourselves with hard facts on claims volumes, Council neglect of roads, plus the grim reality of getting anything fixed in Britain today.
UNIVERSITY STUDY ON AI FRAUD DETECTION
Sticking with fraud, the FT reports on a study by the University of Munster that AI algorithms can detect suspicious patterns within claims books. This can hardly be breaking news to insurtech software creators, or the big insurers who partner with them. The story does make one very good point however; there is an opportunity via AI – real intelligence we mean, not machine learning by rote or simply compressing data then comparing year-on-year stats – for spotting new trends in fraud.
“Since it (AI) does not rely on the categorisation of past claims – a task that requires a lot of resources and is prone to errors – it is less susceptible to relearning fraud patterns that are already known by the insurance company, and can thus potentially detect fraud patterns that have remained undetected in the past.”
But is this sentient HAL9000 AI truly capable of taking raw data and seeing patterns that have not existed before within claims? IE doesn’t think that we are there yet, plus politicians are keen to make AI some type of social justice warrior in robot form, so that means data analytics is bound to be skewed via legislation in the future. As ever with insurtech, proof of concept is what counts, everything else is just a slick slide deck.
CROP SCAM HIGHLIGHTS ALLEGED ORGANISED FRAUD
Sometimes an individual sees an opportunity to play the system, but other times a group of people feel they can work together to falsify data, which will benefit everyone in on the deal. One such case was reported recently in India where crop insurance is a vital part of the rural landscape. Tenant farmers stand or fall by one crop, so a weather event can be disaster.
This case involved a group of perhaps 80-plus people, allegedly faking records on who farmed what piece of land.
The detail that stands out from the trial, which is ongoing, is the accusation that the scheme regarding amending the leaseholding details on land parcels, and insurance coverage, could only have worked with the collusion of some people working for insurance agencies. It’s worth following the case for more we think.
IT’S ALL IN THE DATA, MAYBE
One of those many data related stories we love at IE, this one tracks the uninsured driver percentages in Kilmarnock, (the KA postode in the West of Scotland.) It’s a fairly rural area and as the report in the Largs and Millport News notes, the KA area stats might seem bad, since they are the worst in Scotland, yet they lag behind sleepy Llandrindod Wells in Wales.
That got us thinking; what if poorer drivers in rural areas have little choice in driving to work, since public transport is generally patchy, occasionally once a day? Couple that grim economic reality to a lack of visible policing and you have the perfect circumstances for uninsured drivers taking a chance on commuting, shopping trips etc.
Sometimes understanding human nature is the best way to quantify risk. Just saying.
Be the first to comment