There’s little doubt at this point that self-driving cars are the future of the automotive world. Despite that, the technology still needs some work before it’s safely made available to the public. The following are some current concerns that have kept the self-driving cars away of the open market – at least for the time being.
It’s safe to say that, despite the potential readiness of fully autonomous vehicles within the next few years, self-driving cars will likely be preceded by semi-autonomous ones. While not a full leap forward, this will definitely be a step up for drivers looking to catch a few extra Z’s on the way to work.
The current issue, though, is figuring out whether the driver or the car itself should be responsible in certain situations. Developers are trying to determine how the car should notify the passenger when they’re needed to take over a task. If the driver is watching videos or, more importantly, napping, a reliable form of communication between man and machine is needed. The driver must be informed of their needed involvement, and the car must confirm that the driver is ready to take the wheel.
In terms of current concerns that plague self-driving cars, cybersecurity is definitely up there.
A 2015 example comes to mind, when hackers dug their way into a Jeep’s computing system, bringing it to a full stop on a highway by accessing its onboard entertainment system to reach steering and braking components. In this instance, we saw that even conventional vehicles can have their vulnerabilities exploited, potentially leading to a road accident.
Of course, self-driving cars, with their cloud-backed maps and updates, would pose an even greater cybersecurity risk.
The human mind, while prone to lapses in focus, is built in a way that allows us to judge the best possible reaction to an unexpected encounter. As industry observations have shown, interpreting those situations comes as more of a challenge for self-driving cars. For instance, when analyzing how to respond to a traffic officer directing traffic on a red light, a self-driving car’s programming (based on simple rules) won’t always suffice.
Human beings use contextual clues, such as body language, to judge how to respond in unexpected road encounters. However, it’s difficult for a computer to judge, for instance, whether a traffic officer will be standing perfectly still as the car drives by him, or if he’ll be moving in the vehicle’s way.
This point is, of course, complicated. It’s abstract enough for human beings to navigate their way around the ethical dilemmas of driving – it’s a whole other story for machines. If an unexpected situation arises while driving, and a choice of impact location must be made, which option should the car choose?
For instance, should it make a difference whether a jaywalking pedestrian is a 5-year old girl or an elderly woman? Should the fact that they’re jaywalking make a difference to the direction the car swerves in as a response? These are uncomfortable moral matters, but they’ll need to be addressed clearly and directly before drivers – and pedestrians – can fully embrace automation.