How safe are self-driving or autonomous cars?

The cars of today and the future are getting and continuing to get smarter and smarter, they come with technology that not only helps prevent crashes but also helps take a bit of the responsibility of driving away from the driver. Features such as advanced cruise control technology and parking assist are prime examples of this and are slowly becoming a standard feature in most vehicles.

Manufacturers have also stepped up their game when it comes to safety technology which almost lets the vehicle drive itself. On the more mainstream front, you have examples such as the Nissan Intelligent Mobility safety suite, Toyota Safety Sense, Honda Sensing, and Ford’s Propilot system. While these aren’t exactly autonomous driving technologies it is slowly evolving into it as each new version of it gets smarter and smarter.

On the autonomous driving front, you have automakers like Tesla which already have a functioning system. While the Telsa autopilot system is far from perfect, it’s still a major leap towards fully autonomous driving vehicles. With that said the question now arises, how safe are self-driving or autonomous driving cars?

Before we can answer the question, we must first take a look at the dangers that self-driving vehicles might pose to the road users of today. While technology is indeed evolving at a rapid pace, having a self-driving vehicle might give passengers a false sense of security. If we begin to heavily rely on the computer doing all the driving behind the scenes, we might not be alert enough to take control of the vehicle when a situation happens that is beyond its programming. Regardless even if these vehicles are smart enough to drive on their own, the human driver behind the wheel still needs to be able to take over at a moment’s notice and should always be alert.

A recent example of this was a fatal crash that involved a Tesla vehicle. While the results of the investigation showed that the autopilot was off, CEO of the company Elon Musk has already made moves to phase out the old radar-based autopilot system. Tesla will now be shipping its vehicles with a camera-based autopilot system to further prevent accidents from happening. While having an autopilot system is a great feature to have, it is still best to remain vigilant behind the wheel no matter how good or how advanced the vehicle’s autonomous driving and safety system are.

Just like with any form of computerized technology the threat of a cyberattack is a very real possibility. This has been proven before even with a non-autonomous driving vehicle as hackers were able to take control of a Jeep in 2015. They utilized the vehicle’s internet connection via its entertainment to gain control of the steering as well as the brakes of the vehicle. This poses a risk to future and current self-driving cars as they too could be vulnerable to these kinds of attacks. 

While this may seem like a downside to future self-driving cars, it is preventable with regular security patches by the manufacturers. The more complex a vehicle becomes, the higher the need is to keep everything in check. Only time will tell how manufacturers will be able to deal with this possible threat. One thing that they can do is equip these future autonomous vehicles with an emergency function that surrenders all control back to the human and shuts off any internet or outside connection, the vehicle might have. This will ensure that there is no entry point in the vehicle that would-be hackers might try and abuse.

When it comes to these situations self-driving vehicles are limited by their programming. If a variable is introduced that the system isn’t familiar with, it might end up doing the wrong thing resulting in a crash. This is why it is still better to have a human behind the controls when an emergency situation arises. However, as autonomous driving technology is still evolving the systems could get smarter over time, but as it currently stands the programing and technology behind self-driving cars is still limited and is good for predictable driving conditions or highway cruising use. 

In line with complex driving situations, the artificial intelligence used in vehicles in the self-driving vehicles of today is still limited by their programming. Given a complex emergency situation in which an ethical decision will need to be made, the autonomous vehicle might not make the right decision. 

An example of this would be an emergency situation in which the vehicle will have to choose between saving the driver or harming pedestrians. If its directive is to save the driver and their fellow passengers, then the autonomous vehicle may choose to hit the pedestrians instead. However, if the programming dictates that it should prevent harm to pedestrians then the occupants of the vehicle could be the ones in danger. With many variables to try and account for, this is something that a computer might not be able to comprehend just yet.

As it currently stands, these autonomous cars are not yet perfect but are safe enough for limited use, they can only do what they were programmed to do and can’t really think for themselves. In an emergency situation, it is still better to have a human behind the controls as we can see and account for more variables that the sensors of the vehicle can’t comprehend. We can make decisions with a far better than that of a computer especially comes to ethical choices in which another person’s life is on the line. 

While we can rely on the computer to drive the vehicle safely from point A to point B, we must always be ready to take control at a moment’s notice. However, given how faster technology advances, it is only a matter of time before we can call self-driving or autonomous vehicles safe. Car companies are already developing this technology and are already progressing at an impressive rate, making sure that all future autonomous vehicles will also, to some degree, think and act like a human would in the event of an emergency situation.