Automotive Data Privacy
How Anonymous Data Can Help Americans Accept Connected and Self-driving Cars
Connected cars and advanced driver assistance systems (ADAS) will lower the number of accidents and incidents, forcing insurers to rebuild mechanisms currently used to measure risk and generate revenue. Local and state governments will also feel the pinch of safer driving, through declining revenue from speed and other moving violations.
The general thinking is that as technology increases safety, insurers will shift the search for premium revenue to automotive manufacturers from drivers. Accidents become product liability claims. But I wonder if insurance companies can maintain their stock prices in this brave new world. From the link above:
According to forecasts released by KPMG this week, a decline in accident frequency due to safer vehicles and the adoption of autonomous vehicles could shrink the US personal auto insurance sector by 60% within 25 years.
For FY 2014, Progressive Insurance alone took in $18.4 billion in gross revenue from premiums. What is the amount that the entire US auto industry insurance would need in order to stay at current levels? $200 billion? $750 billion? $2 trillion?
I’m crazy about the benefits of what connected cars and ADAS can do. I’m not crazy about what tinpot government and insurance company dictators could do with vehicle information. I am certain that insurers do not want to be your co-pilot. But record low stock prices and revenue collections have a way of making organizations think differently. I think a knee-jerk response would see them try to exploit the technology to find lost revenue. The best way to accomplish that is to create schemes that make people feel comfortable with exchanging vehicle and personal information for lower than normal fines and premiums.
The scheme is already underway in the insurance industry and is called “usage-based insurance (UBI)”. Over a million Americans are holding their nose and plugging insurance company OBD-II data collectors into their cars, exchanging privacy for lower rates. There hasn’t been a stampede to adopt UBI. This might be because people don’t want insurers as back seat drivers, and vice-versa. Or because insurers have yet to market an offer that people can’t refuse. Either way, self-reporting should continue to grow.
It makes sense that the detail provided from self-reported data would help establish a new baseline for traffic safety. However, as insurers get comfortable integrating streams of almost real-time traffic data into their risk calculations, the question is whether consumers would benefit from the new baseline or whether it is a new starting point for existing rates. I am sure we will never know. But we will pay.
No matter, self-reporting sets consumers up for a new kind of rate increase. If you choose to enroll in these usage-based insurance plans, you plug an OBD II dongle in for a month and an insurance company bases your custom premium on a set of factors. There is nothing keeping them from asking you to plug in again so they can add new factors going forward. The “snapshot” becomes a “movie”, where you’d ultimately leave the dongle in place. There is also nothing keeping them from manipulating your custom rate based on your driving behaviors and how they relate to the baseline. If your average speed is 26 mph, and the new baseline is 22 mph, say hello to a possible increase. If you pass a lot and the baseline says others do not, get out your wallet.
How else could we be tracked and charged? Let us start with a common example, tollway information. My state, the State of Illinois, collects tolls through a system called I-Pass. Your state probably has something similar. A car equipped with a small mobile transponder unit communicates with a fixed reader as the two pass each other. Readers are located at tollway entrances, exits and “open road tolling” areas where masses of them are suspended above a portion of roadway capturing toll information at whatever speed the driver is doing. In a future where governments receive 50% less revenue from motor vehicle penalties, I can see an intrepid politician or two take up the cause of using the system to ticket and fine the vehicle owner.
Prior to building open road tolling stations, Illinois sidestepped an tsunami of outrage when they legislated that readers would not capture the speed of I-Pass equipped cars.
I-Pass is in a category called V2I, or “vehicle to infrastructure”. V2I could be used for great purposes, like giving you proximity to twisty turns, railroad crossings, or public safety vehicles stopped at the roadside. Reflective V2I devices embedded into the centerline of a road could increase the quality of lane markings—-as well as be a way to track VINs and fine you for passing illegally. This information would naturally be forwarded to your insurer, meaning higher rates.
V2V is for “vehicle to vehicle”, and is another category that can be used for good or folly. Under ADAS, cars would transmit data containing speed, direction, operating condition and other key functions to approaching and passing vehicles, significantly lessening the chance of impact. If an oncoming car could tell you that it’s brakes are out, then you could be warned ahead of time. V2V could make for some really ingenious ways to save lives. V2V could also be used to identify the behavior and VINs of vehicles around you, match it with insurance company data and become an indirect way to penalize or reward drivers. As well as fine the vehicle owner.
The solution to the problems above is anonymous and aggregated data.
During the routine exchange of information, whether V2I, V2V, or V2 whatever, the VIN must stay close to anonymous.
Anonymized data means we continue to follow the current convention of investigating accidents and incidents. If fault is found, only then would the VIN and owner identity be revealed to the government and insurance companies. If we need to track a bad guy based on travel patterns or activity, a court order could unlock that information in the same way as it does today.
Aggregating the data means the government would pass on behavioral information without specifically identifying a VIN. For example, the government uses V2I to measure 60 minutes of driving behavior at an intersection. In that hour, 5000 vehicles pass through an intersection and exhibit a number of behaviors. Most normal, some horrible, a few dangerous. Aggregation would conceal the VINs of the individual vehicles and still allow for the measurement of behavior.
The adoption of anonymized and aggregated data means the insurance benefits of real-time traffic information can be spread across all drivers, whether they are in a 1996 Pontiac Sunfire, or a 2017 Volvo S90 with all the semi autonomous bells and whistles. The market experience is unchanged for the consumer, but much more reality-based for the insurer. Consumers may not need to self-report to get a better rate, as the risk profile is much more accurate, and could be refreshed quickly.
The road to mass adoption of self-driving cars is a long time away. The age of semi-autonomous cars, where we have the choice of when to drive or be driven, is upon us. Connectivity and data privacy are important issues that will drive the adoption rate. No one wants an insurance company looking over their shoulder while they drive. Nor do they want to self-report themselves into violations or fines. Legislated rules around data anonymity and aggregation are the only way citizens will understand that they are the ones doing the driving.