I remember watching an episode from one of my favourite detective shows. This one stuck with me more than others because a car equipped with a highly-developed AI engine was hacked while the driver was inside.
As a result, the driver lost all control, the entire thing ended in a nasty car crash, and the driver didn’t make it.
But, even though we still consider unfastened seat belts, drunk drivers, and illegal number plates among the greatest issues for modern drivers, autonomous cars are close to becoming a reality, and the problem of cybersecurity is quite pressing.
And while we do have access to a large selection of legal number plates, cyber crimes are hard to prevent and prosecute. This is especially true if we consider the wave of cyber-crime that has put the world on fire in the last few years.
Everyone was affected, starting with the small guys who don’t really care about IT security and up to the big companies where there is a large annual budget focused on data protection.
So how can we ensure the above-case scenario doesn’t happen in real life? After all, once we completely surrender vehicle controls to AI, we won’t be able to activate the manual switch. And even if we could, in a future where cars drive, the probability of humans learning how to drive is very small.
Learn from Hackers
The good news is that tech companies understand the role of hackers in creating a safe product. While not everyone agrees with the practice, most companies ask white hat hackers to test their systems. These are the good guys of hacking, the ones that work as security specialists, with a sweet tooth for creating chaos in order to identify the weaknesses of a new piece of technology.
So, all safe-driving vehicles will be thoroughly inspected and tested against hacking before they can roll on the streets.
The Tech Is Obfuscated and Complex
A hacker needs to know how a system works in order to discover and exploit its weaknesses. For instance, in the notorious case of the Jeep hacking, it was rather easy to understand how the system worked and how the computer processed the information.
However, if the tech we have right now stands in a few sensors, a self-driving vehicle needs a whole lot more to function properly. The technology will be well-hidden under the frame and there will be a lot more sensors to full. Furthermore, because most sensors have limitations, the system doesn’t trust the findings of just one, it needs other sensors to agree on the same problem before it sends a command to the brakes or gas pedal. In a way, the AI of an autonomous vehicle thinks almost like humans do when they decide something isn’t right. As a result, a truly autonomous vehicle will be more difficult to understand and hack.
The reality of autonomous cars is intertwined with the reality of cyber-security threats, whether we like it or not. Sadly, it seems that most companies don’t take the threat too seriously. As a result, what should be the beginning of a new era of transportation may as well be the end of self-driving vehicles.
If we’re not careful and we don’t discuss this as being one of the most pressing issues of the future, all the work invested by companies until today may be in vain. But, I guess only time will tell on which side of the history we’ll be.
Flickr: Self-driving bus line opening in Tallinn