Keen for insurance priced according to how you drive?

The world of car insurance is in for an overhaul in more ways than one. The hardware in electric vehicles often requires specific repair programmes, which could impact the way insurers deal with them in the long run. And then there’s all the advancements in live telemetry and data logging in just about every new car. 

The latter has been seen as a bit of an opportunity for insurance companies. Several brands have rolled out schemes around the world (known as ‘telematics’ in the industry) that price insurance according to a driver’s habits — how often they break speed limits, how harshly they stop for junctions, how erratic or relaxed they are behind the wheel according to the data. 

Some of these systems are more agricultural than others. In late 2020, Tower Insurance debuted an app called GoCarma, which tracked a person’s driving through their phone’s Bluetooth — collating braking and acceleration data, and how often users would use their phones while driving. Those with a good score would then be able to get a discount off their insurance. 

Tesla has a similar scheme, albeit one which uses a lot more metrics generated by the car itself. It launched the scheme in 2019, although only in five different American states. It works in a fundamentally similar way to the Tower app, but with much more depth. Roll-out has been stifled by American regulators and barriers around privacy. 

Now General Motors is keen to follow suit. Overnight it confirmed plans to launch its own behaviour- and algorithm-based insurance product for GM customers in Arizona, Illinois, and Michigan, with the aim of getting regulator approval by March. 

“I’m very excited about what we have in market and its ability for us to show benefit to the insurance equation,” said Andrew Rose, president of OnStar Insurance and GM’s vice president of global innovation. He added that GM aims to roll out the scheme in a dozen more states in the medium term.

GM says that its telematic system will monitor brake applications, accelerator applications, steering inputs, and things like how often drivers use their seat belts. In future, the system is also expected to use a car’s sensor data for even more of a painted picture, as well as potentially one day monitoring a driver’s head and eye movement. 

Curiously, GM also says that those caught out driving poorly according to their metrics won’t be facing any surcharges early on — although this may change in the future. 

One of the frequently asked questions of these incentives is how these systems can distinguish whether certain sudden inputs done in avoidance of danger aren’t bad behaviours. If a driver has to swerve to avoid a pedestrian suddenly walking in front of them, or slam on the brakes to let an ambulance cross an intersection, will the system acknowledge this as a good behaviour trait or a bad one? 

Tesla’s recent much debated push to give its drivers a rating out of one hundred before granting them access to Full Self Driving could be considered a window into this question. Drivers would often report that any sudden stops, even those justified, would result in their driver safety rating slipping downwards.