The app, under development by Ram Dantu at the University of North Texas, uses the phone’s sensor suite to detect both driving conditions and driver errors. If a driver using the app makes a mistake,
The app then issues a verbal alert, such as “Sudden accelerate”; “Hard braking”; “vehicle wandering detected”; “tailgating detected”; “lane hopping detected”; “bad right (or left) lane change”; or “left (or right) swerve detected”, among other things. It also will warn you not to talk or text.
While early testing is being sponsored by insurers as a way to reward safe drivers, I wonder if the outcome on the horizon is using large datasets of driving records as a much better indicator of risk than claims and collisions. This would be especially valuable among young drivers, who insurers have little information about and must treaty as inherently risky. While smartphone data couldn’t capture all aspects of driver safety (checking your mirrors before changing lanes, for example) it could help suss out risk much earlier in a person’s driving career. Needless to say, it could also provide valuable forensic data for correctly distributing blame from collisions where he-said-she-said testimony would otherwise be all judges or insurance investigators have to go on.
While it’s easy to envision this app being used punitively, it might also be valuable for improving skills and coaching new drivers. In the United States, instruction all takes place when a new driver is first starting out, perhaps in just a few behind-the-wheel sessions. Licensing is a done deal once a driver passes their road test, and a driver might not have a reason to get instruction or coaching on better driving for four or five decades, when insurance companies begin to give discounts for senior refresher classes. A licensing system that requires new drivers to take annual or semi-annual coaching sessions, aimed at improving skills or maneuvers that app data indicates are the driver’s weaknesses or greatest risk factors. If the app records an older driver’s skills deteriorating in a particular area, an insurer could also prompt the driver with an incentive to take a coaching session on improving that skill.
These interventions will inevitably be seen as intrusive, but the fact is that people are terrible judges of their own driving skills. Since the 1980s, a series of studies have consistently found an optimism bias in drivers’ evaluations of their own ability. Clearly, most of us could use a dose of humility when it comes to our skills behind the wheel.
Sooner or later–probably sooner than many of us think!–autonomous vehicles will begin to disrupt our standards of safe driving; Google’s driverless fleet has driven over 300,000 miles accident free, and autonomous vehicles have the potential not just to match but to greatly exceed human safety records. As we navigate the transition between human and automated driving, I expect that insurers and society will raise the bar on acceptably risky driving, and monitoring systems like this one will play a role in allocating and pricing the risks we all impose on each other when we drive.