Late last month, news broke of the first fatal car accident involving a vehicle in “self-driving” mode. The incident, which occurred during a test run, represents a significant blow to the technology, billed as a safer alternative to vehicles operated by human drivers. “Driverless” cars, as they are called, work through a combination of advanced computer control systems, using information provided by GPS, radar, cameras and other means to navigate drivers to their destinations
The incident has been described as a “wake-up call” by experts, and a federal investigation has been launched in the United States, where the accident occurred. But while the crash has been highly publicised and has raised serious doubts as to the readiness of this technology, it is useful to bear a couple of facts in mind.
Firstly, despite being called “driverless” or “autonomous” cars, the current technology is not designed to relieve drivers completely of their responsibilities. Drivers are still expected to keep their hands on the wheel at all times and to take over where necessary. In this, driverless cars provide something more akin to an advanced form of cruise control than a fully autonomous system. We remain a long distance from technology that would allow drivers, say, to take a nap or read a book while their cars drove themselves. The crash in this case is thought to have been caused by a white truck crossing the road ahead of the car, which could have blended with the sky and fooled both the car’s computer systems and its human driver.
Further, though the sample size is inevitably small at this stage, driverless cars are statistically safer than normal cars. In 2014, the National Highway Traffic Safety Administration (NHTSA) recorded thatone person dies for every 100 million miles driven in the United States. This fatality, on the other hand, was the first in 130 million miles driven by driverless cars. Interestingly, NHTSA has recorded that “human error” is the cause of 94% of all accidents on the road.
It is also important to understand that the technology will only become safer the more ubiquitous it becomes. This will happen in the form of incremental refinement, but more importantly, when all cars on the road are driverless, this will allow for the development of a cohesive, coordinated system. While human drivers have only their indicators and horns to communicate with, driverless cars will be able to communicate wirelessly with all the other cars in the vicinity. They will thus form a kind of moving network, and will have a far more accurate awareness of where other cars are going and will know further in advance, for instance, when a car is changing lanes on the motorway.
As companies like Google and General Motors continue to invest heavily in this technology, the proliferation of driverless cars seems inevitable, and mass market use seems more a question of “when” than “if.”
These developments will no doubt be followed closely by insurers in the years to come, as driverless cars have the potential to be immensely disruptive to the industry. In PwC’s 2016 Global CEO Survey, it was found that “insurance CEOs rank behind only the entertainment and media sector in their levels of concern about growth prospects,” with emergent technologies like driverless cars being cited as a key reason behind this, as they herald “fundamental changes in the types of risk that the industry is insuring against.” In an interview for the survey, Craig Olsen, CEO of IAG New Zealand, specifically pointed to driverless cars as a particular threat to that company’s growth. Experts from within the industry have predicted a dramatic reduction in premiums due to these potential improvements in road safety, expecting them to fall as much as 50% by 2025.
It is important, however, that we do not let fear of the new undermine progress that could benefit all of society. If we acknowledge that these changes are coming, the best approach is to start thinking about how we might adapt to meet these challenges.
While emergent technologies may threaten growth in some areas, they equally create new opportunities for growth. Cyber insurance, for instance, is sure to become more relevant and necessary as we become more reliant on computer software in our daily lives. To this, it is worth pointing out that a key risk of driverless cars is the possibility that third parties will be able to hack into their computer systems.
Ultimately, driverless cars are not going to remove the public’s need for car insurance. Cars will still be stolen, broken into and vandalized. What is likely to change, however, is where the liability lies. We would expect liability to shift away from drivers themselves to focus more on other parties. This could mean the cars’ manufacturers, or the developers of their automated computer systems. It could also mean telecommunications providers, if signal failure is the cause of an accident. The driver, of course, can still be at fault if the situation called for some manual override. Fundamentally, liability will continue to hinge on the facts of each individual incident, as has always been the case. We simply need to familiarise ourselves with different sets of facts from those to which we have become accustomed.