The questions surrounding driverless vehicles for insurance underwriters, GI companies and brokers are huge.
Who is liable when an accident occurs; the council for poor road signage and markings, the camera software supplier, the LIDAR fitment & service workshop, or was the AI program that effectively assimilates all that data and `drives’ the car also partially to blame?
Yeah, it’s more complicated than Brexit: The Musical. So Insurance Edge chatted with Tim Loo, who has been working on an interesting project with AXA XL.
IE: Let’s get straight into the blame aspect, because that tipping point of shifting liability from the driver, to the software, as autonomous vehicles develop, is probably going to be the most difficult thing to decide for insurers?
TL: It certainly is, but all this stuff is being built so the industry as a whole has to jump in at some point and start to work out common standards, or perhaps help develop those rules with government agencies and regulators.
There are already lots of companies involved in creating the driverless vehicle chain. The manufacturers are buying in all the parts, plus all the software from dozens of suppliers, so it’s a big chain of potential liability. For example, the software makers may well be held responsible for any necessary updates to cameras, LIDAR, automatic braking systems etc. But then what if the owner/renter of the car failed to install the software update and there was a crash – where is the blame then?
You know the interesting thing about the switch to driverless is that for many insurers and brokers, maybe even comparison sites too, the business model they have right now will not be workable in the future. I mean if you don’t own the car, you rent it/share it, plus you aren’t physically driving it, then why would you purchase insurance such as Fully Comp or TPF&T? There is an opportunity for insurers to develop new products that fit the future market.
IE: Tell us more about the AXA XL project.
TL: At Foolproof we partner with lots of brands in the mobility space, such as Suzuki, Avis and Shell, but the AXA XL initiative is particularly exciting because this is a case of an insurer taking that leap into the unknown and trying to develop a roadmap for autonomous vehicles, something truly practical. We aren’t interested in simply bouncing around ideas and concepts, but working out a system that can work in the real world, on every street in any city, town and village.
See a longer video presenation by Tim and AXA XL here by the way.
For AXA XL, we wanted to look ahead and try to understand the broad stages of autonomous progress and how insurance has to develop in tandem with those steps. For example, are governments going to ban diesel engines completely? Will there be new laws and driver testing regimes tied into semi-autonomous cars, as the tech improves? How can insurers factor in the costs of protecting autonomous vehicles from cyber-attack(s) because malicious actions are a real risk as the number of driverless vehicles increases.
Then there’s an extra layer here; the car itself is gathering data and effectively becoming the underwriter of the risk, in real time. AI is making sense of the terrain, speed, other vehicles’ proximity, pedestrians, cyclists, road signage, weather conditions and so on. But if that automated risk assessment fails, then the insurance industry needs to create a failsafe position, some kind of back-up. Otherwise people will not have the confidence to invest in driverless – emotionally as well as financially.
I recently travelled in an autonomous car, which has been stationed in Stratford for the last 6 weeks or so, and it is a curious mixture of regular mundane transportation with moments of high-level excitement. I guess everyone is still at the stage of building trust in this new technology and AXA XL, plus their partner Oxbotica, are very much focussed on the safety aspect. As we build this new transport ecosystem, it will benefit everyone if we err on the side of caution.
IE: Oxbotica and AXA XL are pioneering some trials at airports, which is a double-edged sword in that there aren’t variables like random traffic, cyclists, pedestrians etc airside, but you have jet aircraft full of fuel and people. Why choose airports to trial true autonomous vehicles?
TL: Many reasons in fact. Firstly, there is a cost benefit case for the airport industry. You have lots of staff airside moving people on buses, luggage, fuel tankers, maintenance vehicles and so on. Security at airports is a 24/7 business, because you can’t have public access at all. So if you remove the human element from those two functions, you start to save a great deal of money.
Secondly, human beings make mistakes, or are sometimes bad actors for various reasons. By going autonomous, you remove that layer of risk too. The third part of this equation is that airports already have clearly defined tracks and roads, guide lighting on those trackways, defined zones and areas for particular vehicles and so on. You have an ecosystem in place, so all you need to do is install the new technology and enhance the existing location’s functions.
IE: So the lessons we learn from airports can translate into busy cities. Baggage handling becomes food delivery, passenger transport from fuselage to Gate is a dry run for the average commute. Will autonomous vehicles catch on as quickly in rural areas, where people may not travel so often, or just value the independence aspect of owning your own car?
TL: Yes that’s probably going to be how driverless take-up spreads, cities first where the benefits of owning are car versus renting are marginal. But Oxbotica are able to make this tech work anywhere, regardless of the terrain and once there’s infrastructure, we will probably see a tipping point where the take-up is quite sudden.
IE: Kind of like contactless debit cards in the last few years?
TL: Yes, people are wary at first, and then as they see their friends aren’t being defrauded, plus don’t have to wait in a long queue to buy a ticket with cash, or type in a PIN, then they give it a go too.
IE: Do you think that insurers will need to switch to a Pay-per-Mile or PAYG model, taking insurance as a bolt-on to a vehicle service subscription, sort of CarFlix model?
TL: There is a distinct possibility that insurance will be bundled with a hire service yes.
Moreover, insurers will arguably need to start thinking about becoming service providers, crisis management companies if you like, rather than risk assessors. The tech and the data gathered will underwrite much of the risk, but we will need someone to kick into action when things go wrong.
That’s the value for consumers, the back-up. The big change in motor insurance is going to be the shift in liability from driver to software, and the natural consequence of that change is that you won’t need a broker or insurer to rate the driver’s risk by postcode, job, age, convictions etc.
A whole new paradigm of risk associated with developing software systems will probably become the primary focus for insurers. We are some years away from true autonomous travel, but it is great that companies like AXA XL are trying to figure out how to make it work right now.
IE: Tim, fascinating stuff. Thank you.