Telematics vs self-driving cars: No contest
Well thanks for dropping by to read my latest musings on connected insurance. If you would like to subscribe to future issues, please click here.
A few years ago, I was in a meeting with the CTO of a large motor insurer. We were discussing Telematics enabled Usage Based Insurance (UBI). At one point, he turned to me and blithely said “With self-driving cars, your Telematics system will be useless!”. It was a thought provoking comment and one that I’ll address in this month’s column.
Can you remember when it was perfectly legal to drive a car without wearing a seatbelt? Where I come from, that era ended in 1970. It was Victoria, Australia where, for the first time in the world, it became illegal to be in the front seat of a car and not wear a seatbelt.
How far we have come! We can now conceive of a time when it may be illegal for a human to be in the front seat, driving a car on a public road. I wonder where in the world, will be the first place to take that bold step?
Here come the Robots – slowly
Well, hold on there for a moment, Sunshine!
With over 1.2 billion vehicles on the road today, and given that the rate of new vehicle production sits at around 68 million units per year, it’s going to take several decades to replace them all. Additionally, we’re not switching production to 100% self-driving cars tomorrow. Judging by the announcements of the major manufacturers, it will be quite some time before more than a small fraction of those new cars will be completely autonomous.
Volvo is targeting 2017 for its first autonomous vehicle production release, while other vehicle manufacturers are planning for initial production releases in 2020 and some as far out as 2022. We can all agree – the introduction of fully self-driving cars is going to be a phased affair over a very long time, with varying degrees of autonomy introduced at each phase of the transition.
Throughout the history of the automobile, we have witnessed many safety innovations progressively implemented in a similar fashion. Most start out as optional features, but then eventually wind up being mandatory or standard. From the advent of the humble seatbelt, through to airbags and crumple zones on cars, these have all taken many years to become standard. Now we are seeing the same pattern repeat with autonomous driving.
Crash prevention or reduced accident severity
The latest of those innovations is automatic emergency braking (AEB). In the US, auto makers recently agreed to make this safety feature standard on all models by 2022, with a progressive introduction over the next six years. AEB, once fully rolled-out in the US, is expected to reduce the number of accidents by over 9,000 annually.
This development is interesting for a number of reasons. Firstly, according to the US government, making the introduction voluntary gets the technology in place faster than if mandated through regulation. Secondly, the government is using a consumer watchdog organization to police the introduction. Finally, there seems to be no absolute standard around the AEB technology with some models bringing vehicles to a complete stop to avoid an accident while other models simply reduce speed to make an accident less severe.
Robot driving style
This last point is quite telling, in so far as not all AEB systems are created equal. It seems we are now developing robot driving “behaviour” that will vary between make, model, year of manufacture, and potentially also software patch level? It’s a curious notion.
Given there are now over 30 different companies involved in perfecting self-driving cars, it’s very likely we will see a similar situation emerge in autonomous performance. We should expect that each brand will have its own unique set of driving eccentricities – at least initially.
Driving style will be further complicated during this long transition period, with a mixture of human and robot behind the wheel. If the human driver is able to switch on or off autonomous features, then we have “first-line” driving responsibility potentially shifting between human and robot from moment to moment. This is exactly like Teslar’s Autopilot feature, where the driver must have hands on the steering wheel, while the feature is engaged.
Telematics steps in
For human drivers, insurers are able to use telematics to record driving behaviour and predict the risk of an accident. These same telematics systems are quite capable of determining which autonomous safety features are active at any time, just as they are able to know when a driver is in ride-sharing mode for Uber or Grab.
With a telematics based program, we have the opportunity to dynamically re-rate premium based on the observed behaviour. For example, a human driver whom we segment into the “bad driver” (high risk) category, may be transformed into a good or excellent driver (low risk) category, the moment autonomous driving mode is triggered. An insurance telematics program could therefore help provide incentives for drivers to engage and spend more time in self-driving mode rather than driving their vehicles themselves.
Telematics will also enable the systematic collection of robot driving behaviour data whilst autonomous driving is engaged. This will allow insurance actuaries to independently ascertain the risk characteristics across the whole product range of autonomous vehicle manufacturers. Thinking back to the US government’s approach to AEB, I would certainly want insurance actuaries backing up any consumer rights watchdog organization in the determination of the relative safety of self-driving systems.
Looking behind the curtain
Historically, vehicle manufacturers have been reluctant to share any data with telematics system providers under the guise of protecting intellectual property. This veil of secrecy, could pose a problem for the use of insurance telematics with self-driving cars. Fortunately, governments are taking a very proactive and transparent approach to regulating the trialing and introduction autonomous vehicles. The approach should mitigate this risk, especially as regulators, in some cases, are already considering the implications for insurance.
The stakes are high, as governments know there will be (have been already) fatalities while autonomous vehicles are being implemented. But they balance that against the reality of over 1.2 million lives being lost on our roads each year and that over 90% of motor accidents are the result of human error. Autonomous vehicles present a real opportunity to dramatically cut the toll on human life by eliminating those errors. But this can only happen within an open and transparent process, where data is shared.
The bottom line
Telematics and self-driving vehicles have a logical affinity, although it’s not that obvious at first glance. For insurers who wish to survive the transition from human to robot drivers, telematics is an excellent way to start on that journey. There seems little doubt the transition has begun. As an industry, we are in for an exciting ride over the coming decades as self-driving cars become commonplace. The winners will be those insurers who can turn the flood of data becoming available from human and robot drivers into actionable knowledge – knowledge that benefits their customers, the government, the community, vehicle manufacturers and of course the insurers themselves.
BTW: Toot twice if you are a robot.
I hope you had as much fun reading this column as I did writing it. If you did, please let me know by subscribing. I’ll be running some give-aways for loyal subscribers later in the year, so it pays to subscribe! If you want to see more of my “off the wall” comments, you can follow me on Twitter @ITInsuranceGuy. Finally, I am keen to hear what you have to say, so please email me at firstname.lastname@example.org and share your views, insights and stories about technology in the insurance industry. I look forward to interacting with you.
The author, Andrew Dart, is an Insurance thought leader, telematics practitioner and editor of “Insurance Connected” a regular column for The Digital Insurer.